JP7545933B2 - Agricultural support systems and mobile devices - Google Patents

Description

The present invention relates to an agricultural support system and a mobile object.

In recent years, smart agriculture has been introduced in an attempt to reduce the labor required for agricultural work and make it more efficient, reduce the workload of existing agricultural workers, and encourage new farmers to enter the industry and pass on agricultural techniques. Smart agriculture makes use of robotics and information and communications technology to reduce labor required for agricultural work and make it more efficient.

For example, Patent Document 1 discloses a system (hereinafter referred to as the "conventional system") that analyzes data acquired by cameras and sensors placed on farms and agricultural robots on a server and outputs information to support the work of agricultural workers and agricultural robots.

Conventional systems can assist farmers and agricultural robots in agricultural work based on various data about the farm and crops acquired by agricultural robots moving around the farm.

JP 2019-82765 A

However, in conventional systems, the agricultural robot's purpose is to use a camera to capture images of the farm worker's work on the farm, so it is positioned behind the farm worker and moves with the farm worker, for example. This means that when the agricultural robot collects data on the farm or crops, it is assumed that the farm worker is working on the farm, and data cannot be collected while the farm worker is not working, which creates a problem in that the burden on the farm worker cannot be reduced. To solve this problem, there is a demand for an agricultural support system that uses a mobile object that can collect farm data while traveling autonomously around the farm, even when there are no farm workers present.

In farms where such agricultural support systems are applied, for example, multiple rows of ridges and furrow spaces between the ridges are arranged in an alternating fashion. After the mobile body finishes traveling through one furrow, it moves to the adjacent furrow, and repeats this process, traveling autonomously through all the furrows in the farm. However, if the mobile body is unable to turn at an appropriate position after traveling through a furrow and exiting it, the mobile body may come into contact with or ride up onto the furrow, causing part of the ridge to collapse.

The present invention has been made to address the above-mentioned problems. That is, one of the objectives of the present invention is to provide an agricultural support system and a mobile body that can reduce the possibility of part of a ridge collapsing when the mobile body, which travels autonomously within a farm, moves from one furrow to an adjacent furrow.

In order to solve the above problem, the present invention provides an agricultural support system including an information processing device and a mobile body capable of traveling on a farm formed with a plurality of rows of ridges and furrows arranged side by side, the mobile body being configured to be able to communicate with each other, the mobile body being equipped with a communication device capable of communicating with the information processing device, a storage unit in which a farm map corresponding to the farm is stored, the farm map including ridge information including the positions of the plurality of ridges and the furrows, a first information acquisition device that detects objects by acquiring object information related to objects around the mobile body, a second information acquisition device that acquires mobile body information related to the state of the mobile body, a position information acquisition device that acquires self-position information indicating the position of the mobile body, a traveling device capable of traveling the mobile body in a desired traveling direction, and a control device that controls the traveling device so that the mobile body travels autonomously on the furrows based on the farm map, the object information, the mobile body information, and the self-position information, the control device being configured to control the traveling device so that the mobile body travels autonomously on the furrows when the mobile body is traveling on one of the furrows, When the moving body goes from a ridge detection state in which the ridges are detected on the left and right sides of the moving body to a ridge undetected state in which the ridges are not detected on the left and right sides of the moving body, the traveling device is controlled so that the moving body moves in a straight line a first distance from the position at which the first information acquisition device went into the ridge undetected state to a first turning position, turns at the first turning position, and then performs a first straight line movement in which the moving body moves in a straight line along the direction in which the multiple ridges are arranged toward a position where the ridges are present only on one of the left and right sides of the moving body. Furthermore, the control device is configured to control the traveling device so that when the first information acquisition device goes from a state in which the ridges are detected only on one side of the moving body to a ridge undetected state in which the ridges are not detected on the left and right sides of the moving body, the moving body moves straight a second distance from the position at which the ridge undetected state was detected to a second turning position, turns at the second turning position, and then performs a second straight movement operation in which the moving body moves straight toward the furrow adjacent to one of the furrows.

The present invention is an agricultural support system including an information processing device and a mobile body capable of traveling on a farm formed with a plurality of rows of ridges and furrow spaces, the mobile body being configured to be able to communicate with each other, the mobile body being equipped with a communication device capable of communicating with the information processing device, a storage unit in which a farm map corresponding to the farm is stored, the farm map including ridge information including the positions of the plurality of ridges and the furrow spaces, a first information acquisition device for acquiring object information regarding objects around the mobile body, a second information acquisition device for acquiring mobile body information regarding the state of the mobile body, a position information acquisition device for acquiring self-position information indicating the position of the mobile body, a traveling device capable of traveling the mobile body in a desired traveling direction, and a control device for controlling the traveling device so that the mobile body travels autonomously on the furrows based on the farm map, the object information, the mobile body information, and the self-position information, the farm being equipped with a first marker installed at a position a first predetermined distance away from a first turning position and a second predetermined distance away from a second turning position. The control device is configured to: when the moving body is traveling in one of the furrows, the control device acquires a first marker distance, which is the distance between the first marker and the moving body, as the object information by the first information acquisition device; and when it is determined that the moving body has reached the first turning position based on the first marker distance and the first predetermined distance, control the traveling device so that the moving body turns at the first turning position and then moves straight along the direction in which the multiple ridges are lined up toward a position where the ridges are present only on one of the left and right sides of the moving body; and when it is determined that the moving body has reached the second turning position based on the second marker distance and the second predetermined distance, control the traveling device so that the moving body turns at the second turning position and then moves straight toward the furrow adjacent to the one of the furrows.

The present invention relates to a mobile body configured to be capable of communicating with an external information processing device and capable of traveling on a farm formed with a plurality of rows of ridges and furrows, the mobile body comprising a communication device capable of communicating with the information processing device, a storage unit in which a farm map corresponding to the farm is stored, the farm map including ridge information including the positions of the plurality of ridges and furrows, a first information acquisition device that detects objects by acquiring object information related to objects around the mobile body, a second information acquisition device that acquires mobile body information related to the state of the mobile body, a position information acquisition device that acquires position information indicating the position of the mobile body, a traveling device capable of traveling the mobile body in a desired traveling direction, and a control device that controls the traveling device so that the mobile body travels autonomously on the furrows based on the farm map, the object information, the mobile body information, and the self-position information, and the control device detects the ridges on the left and right sides of the mobile body when the mobile body is traveling on one of the furrows. and when the first information acquisition device goes from a ridge detection state in which the first information acquisition device detects a ridge on either the left or right side of the moving body to a ridge undetected state in which the ridges are not detected, the moving body moves in a straight line a first distance from the position at which the first information acquisition device went into the ridge undetected state to a first turning position, turns at the first turning position, and then performs a first straight line movement in which the moving body moves in a straight line along the direction in which the ridges are arranged toward a position where the ridges are present only on one of the left and right sides of the moving body; and further, the control device controls the traveling device so that When the first information acquisition device goes from a state where the ridges are detected only on one side of the moving body to a ridge undetected state where the ridges are not detected on the left and right sides of the moving body, the traveling device is controlled so that the moving body moves in a straight line a second distance from the position at which the ridge undetected state was detected to a second turning position, turns at the second turning position, and then performs a second straight line movement toward the furrow adjacent to one of the furrows.

According to the present invention, when a mobile object traveling autonomously within a farm moves from one furrow to an adjacent furrow, the possibility of part of the furrow collapsing can be reduced.

FIG. 1 is a schematic diagram showing an example of the configuration of an agricultural support system according to a first embodiment of the present invention. FIG. 2 is a functional block diagram showing an example of the configuration of a robot. FIG. 3 is a perspective view showing an example of the configuration of a robot. FIG. 4 is a plan view showing an example of a farm on which a robot travels. FIG. 5 is a schematic diagram showing the robot moving between furrows in a farm. FIG. 6 is a diagram for explaining the relationship between the monitoring images of the furrows and the monitoring images stored in the server. FIG. 7 is a diagram for explaining the relationship between the monitoring image of the furrows and the monitoring image stored in the server. FIG. 8 is a diagram for explaining the relationship between the monitoring images of the furrows and the monitoring images stored in the server. FIG. 9 is a perspective view showing an example of the configuration of a robot. FIG. 10 is a schematic diagram showing the robot moving between furrows in a farm. FIG. 11 is a schematic diagram showing how the robot moves between furrows while detecting the side of the furrows with an external sensor. FIG. 12A is a diagram for explaining the mounting structure of an external sensor. FIG. 12B is a diagram for explaining the mounting structure of the external sensor. FIG. 12C is a diagram showing an example of the configuration of a robot in which the mounting structure of an external sensor is made movable. FIG. 13 is a diagram showing the state of laser light irradiation when the robot shown in FIG. 12C travels between furrows. FIG. 14 is a diagram showing the state of laser light irradiation when the robot shown in FIG. 12C travels between furrows. FIG. 15 shows the state of the robot when it cannot move forward due to the mud between the furrows. FIG. 16 is a flow chart for explaining the operation of the robot when the robot escapes from the mud. FIG. 17 is a functional block diagram showing an example of the configuration of a robot. FIG. 18 is a table for explaining various sensors. FIG. 19 is a diagram for explaining an example of the operation of the robot when moving between adjacent furrows. FIG. 20 is a flow chart for explaining the operation of the robot when the robot moves between adjacent furrows. FIG. 21 is a functional block diagram showing an example of the configuration of a robot. FIG. 22 is a diagram for explaining an example of the operation of the robot when moving between adjacent furrows. FIG. 23 is a flow chart for explaining the operation of the robot when the robot moves between adjacent furrows. FIG. 24 is a diagram for explaining an example of the movement of the robot between furrows. FIG. 25 is a flow chart for explaining the operation of the robot when it moves between adjacent furrows.

<<First embodiment>>
Hereinafter, an agricultural support system S according to a first embodiment of the present invention will be described with reference to the drawings.

<Configuration>
Fig. 1 is a schematic diagram showing a configuration example of an agriculture support system S. As shown in Fig. 1, the agriculture support system S includes a robot 101, a server 102, and a terminal 103 that are connected via a network N. The server 102 may be either a cloud server or an on-premise server. Furthermore, the server 102 may be composed of multiple servers.

The server 102 includes, for example, a computer including a CPU, ROM, RAM, and an interface I/F, as well as a storage device capable of writing and reading data. For convenience, the server 102 may also be referred to as an "information processing device." The terminal 103 is, for example, an information terminal such as a personal computer, a smartphone, or a tablet terminal.

The charging station 104 is installed in the farm 118 (see FIG. 4 described below). The charging station 104 supplies power via a wired or wireless connection to the robot 101 waiting at the charging station 104 and/or traveling within the farm 118.

The robot 101 is a mobile object that performs monitoring work of the farm 118 by autonomously traveling (moving autonomously) within the farm 118 in cooperation with the server 102 and the terminal 103. The robot 101 functions as an agricultural support robot for supporting agriculture. FIG. 2 is a functional block diagram showing an example of the configuration of the robot 101. FIG. 3 is a perspective view showing the external appearance of the robot 101.

As shown in Figures 2 and 3, the robot 101 includes a control device 105, a monitoring device 106, an external sensor 107, an internal sensor 108, a communication unit 109, an antenna 110, a GNSS (Global Navigation Satellite System) unit 111, a motor 112, a pair of drive shafts 113, a pair of drive wheels 114, a power supply unit 115, and a battery 116. For convenience, the external sensor 107 may also be referred to as the "first information acquisition device." For convenience, the internal sensor 108 may also be referred to as the "second information acquisition device."

The control device 105 is an arithmetic processing device such as a microprocessor, and performs overall control of the robot 101 (control of the operation of the robot 101). The control device 105 includes, for example, a CPU, ROM, RAM, and an interface I/F. The control device 105 includes a timer 105a. The timer 105a is used to measure time and/or duration and to execute interrupt processing.

The control device 105 is connected to the monitoring device 106, the external sensor 107, and the internal sensor 108. The control device 105 is configured to receive data acquired by each of the monitoring devices 106 and each sensor.

The monitoring device 106 is an imaging device (e.g., a camera). The monitoring device 106 captures an image of a predetermined range around the robot 101 (e.g., a predetermined range in front of the robot 101) to obtain a monitoring image (monitoring data) that is an image of the surroundings of the robot 101, and transmits the image to the control device 105.

The external sensor 107 is a sensor for acquiring information (also referred to as "external sensor data") outside the robot 101. The external sensor 107 acquires information outside the robot 101 and transmits it to the control device 105.

The external sensor 107 is typically a depth sensor (Light Detection And Ranging (LiDAR). The external sensor 107 may also be a Laser Detection And Ranging (LADAR) sensor, a Time of Flight (ToF), a stereo camera, an ultrasonic sensor, a proximity sensor (infrared sensor, magnetic sensor, etc.), a thermal sensor (e.g., a far-infrared sensor), a temperature sensor, a humidity sensor, a gas sensor that detects a gas concentration (e.g., _CO2 concentration), an illuminance sensor, etc.

The internal sensor 108 is a sensor for acquiring information (sometimes referred to as "internal sensor data") about the state of the robot 101 used to control the autonomous movement of the robot 101. The internal sensor 108 includes, for example, a gyro sensor, an acceleration sensor, a magnetic sensor, a temperature sensor, a power supply voltage sensor, and a battery remaining capacity sensor. The internal sensor data includes, for example, the orientation of the robot 101, the inclination (tilt angle) of the robot 101, the acceleration of the robot 101, the speed of the robot 101, the rotation angle (turning angle) of the robot 101, and geomagnetism. The internal sensor 108 acquires the internal sensor data and transmits it to the control device 105.

The control device 105 is further connected to a communication unit 109 and a GNSS unit 111. The communication unit 109 is a communication interface that wirelessly communicates with an external computer (such as the server 102) via an antenna 110, and that allows the GNSS unit 111 to communicate with other robots and base stations in order to estimate (obtain) its own position information.

The GNSS unit 111 acquires self-location information indicating the current location of the robot 101 from a positioning satellite, and transmits the acquired self-location information to the control device 105. For convenience, the GNSS unit 111 may also be referred to as a "location information acquisition device."

The control device 105 determines the current position of the robot 101 based on the self-position information. Furthermore, the control device 105 determines the relative position with respect to the ridges (e.g., the relative position with respect to the ridges in the width direction between the ridges) based on the external sensor data and the internal sensor data.

The control device 105 is further connected to two motors 112. Each of the two motors 112 generates torque to rotate (drive) the drive shaft 113 corresponding to the respective motor 112 by power supplied from the power supply unit 115 to the motor 112. The motor 112 can rotate the drive wheel 114 corresponding to the respective drive shaft 113 by this torque. The control device 105 can freely control the rotation speed and rotation direction of the drive wheel 114 by sending a command to the motor 112 and controlling the power supplied to the motor 112 to adjust the magnitude and direction of the torque generated by the motor 112.

As shown in FIG. 3, the surfaces 114a of the left and right drive wheels 114 have a characteristic shape (for example, a shape in which convex and concave portions each having a V-shaped planar shape are repeated). This allows the robot 101 to more easily crush grass present in the path of the drive wheels 114. In order to increase the contact area of the left and right drive wheels 114 with the ground, it is preferable that the width of the drive wheels 114 is set wider. From the perspective of crushing grass, it is preferable that the drive wheels 114 have a weight that allows them to crush grass with their own weight.

The power supply unit 115 supplies the power stored in the battery 116 to the control device 105 and each part of the robot 101 in order to operate the robot 101. The battery 116 stores the power supplied from the charging station 104 (see FIG. 1) via a wired or wireless connection.

The control device 105 controls the motor 112 to drive the drive shaft 113, thereby controlling the movement of the robot 101. For example, when the robot 101 moves straight forward, the control device 105 drives the drive shaft 113 so that the left drive wheel 114 (hereinafter also referred to as the "left wheel") and the right drive wheel 114 (hereinafter also referred to as the "right wheel") rotate at the same speed, for example, in a forward rotation direction corresponding to the direction in which the robot 101 moves forward.

For example, when the robot 101 moves straight backwards, the control device 105 drives the drive shaft 113 so that the left and right wheels rotate at the same speed in a backward rotation direction corresponding to the direction in which the robot 101 moves backwards.

For example, when the robot 101 turns in a right direction, the control device 105 drives the motor 112 so that the left wheel becomes faster than the right wheel and the left and right wheels rotate in the same forward rotation direction. For example, when the robot 101 turns in a left direction, the control device 105 drives the motor 112 so that the right wheel becomes faster than the left wheel and the left and right wheels rotate in the same forward rotation direction. When the robot 101 changes its direction of travel on the spot, the control device 105 drives the motor 112 so that the right and left wheels rotate in opposite directions at the same speed.

In this way, the control device 105 can freely control the movement (travel) of the robot 101 by controlling the motor 112. That is, the control device 105 can freely control the movement of the robot 101 to move forward, backward, turn, curve, adjust the direction of travel, avoid obstacles, stop, etc. For convenience, the motor 112, drive shaft 113, and drive wheels 114 may also be referred to as the "travel device."

<Robot operation 1 (monitoring operation (data collection), weed control operation, obstacle avoidance operation)>
The robot 101 uses the monitoring device 106 to capture images of the surroundings of the robot 101 while autonomously traveling in the farm 118, thereby acquiring monitoring images and thereby monitoring the farm 118 and the crops. Fig. 4 is a plan view of the farm 118 on which the robot 101 travels. Fig. 5 is a schematic diagram showing the robot 101 traveling in the furrows 120 of the farm 118.

In the examples of Figures 4 and 5, one robot 101 is responsible for monitoring one farm 118. However, this is not limited to the above, and multiple robots 101 may be responsible for monitoring a single farm 118 by dividing the area. Alternatively, multiple robots 101 with different collected data or functions may be responsible for monitoring a single farm 118.

The farm 118 has ridges 119 and furrow spaces 120 in which crops are planted. In this specification, the space between two adjacent rows of ridges 119 is referred to as furrow space 120.

Although not shown in FIG. 5, farm 118 may not have ridges 119 but may have paths between crops, such as an orchard. In this specification, in this case, the area where crops exist corresponds to ridge 119, and the paths between crops correspond to furrow space 120.

Furthermore, although not shown in Figure 5, the ridges 119 may have bare soil or may be covered with mulch sheets, nonwoven fabric, cut plants, etc., for the purpose of protecting the crops and soil.

A charging station 104 is installed on the farm 118, which serves as the home position of the robot 101.

The robot 101 and the external server 102 store a plan view of the farm 118, such as that shown in FIG. 4, created in advance as map information (hereinafter referred to as a "farm map"). For example, the farm map is stored in the ROM of the control device 105 of the robot 101. The farm map is also stored in a storage device provided in the server 102. When creating a farm map, when creating ridges 119 in the farm 118, the positions (ridge information) of the ridges 119 and furrow spaces 120 are measured, for example, by tracing the position of the tractor, measuring with a drone after creating the ridges 119, or running the robot 101 remotely. Then, the farm map is created based on these measurement results. In the farm map, the positions of the ridges 119 and furrow spaces 120 are associated with position information (e.g., latitude, longitude) on the farm map. Therefore, each of the robot 101 and the server 102 can identify the positions of the furrows 119 and the furrow spaces 120 within the farm based on the farm map.

The robot 101 autonomously travels within the farm 118 starting from the charging station 104, traveling along a route including the furrows 120 indicated by the arrow in FIG. 4 (hereinafter referred to as the "autonomous travel route"), and returning to the charging station 104 again.

Such autonomous driving of the robot 101 can be achieved by the control device 105 controlling the motor 112 based on at least one of the farm map, external sensor data, internal sensor data, and self-position information.

For example, the robot 101 sets an autonomous driving route based on a farm map, and drives along the autonomous driving route based on its own position information.

Furthermore, while the robot 101 is autonomously traveling, it uses the external sensor data, the internal sensor data, and its own position information to travel between the crops as an aisle so as not to run over the crops, and it uses the furrows 120 as an aisle so as not to run over the ridges 119.

While the robot 101 is autonomously traveling, it uses the monitoring device 106 and the external sensor 107 to acquire (collect) data about the farm 118, and transmits the acquired data to the server 102. Upon acquiring the data, the server 102 processes and manages (stores) the data.

The following describes in more detail the data processing and transmission of the robot 101 and the data processing of the server 102 with reference to FIG. 6. FIG. 6 is a diagram for explaining the relationship between the monitoring image 121 of the furrow 120 and the monitoring image 121 stored in the server 102.

While the robot 101 is traveling autonomously, the robot 101 acquires a monitoring image 121 of the furrow 120 by the monitoring device 106. Note that the monitoring image 121 of the furrow 120 in FIG. 6 shows the furrow 120 through which the robot 101 is traveling, the furrows 119 on both sides of the furrow 120, and obstacles such as weeds 122 present in the furrow 120.

The robot 101 associates the acquired monitoring image 121 of the furrow 120 with the time at which the monitoring image 121 of the furrow 120 was acquired (the time acquired by the timer 105a) and its own position information (its own position information acquired by the GNSS unit 111, etc.). The robot 101 transmits a data set of the associated monitoring image 121 of the furrow 120, its own position information, and the time to the server 102 via the communication unit 109.

As described above, the server 102 has farm map data 123, which is a farm map. When the server 102 receives a data set from the robot 101, the server 102 associates and stores the monitoring image 121 of the furrow 120, the acquisition time of the monitoring image 121 of the furrow 120, and the location information in the farm map data 123 that corresponds to the robot's own location information. This allows the server 102 to identify where and when within the farm 118 the monitoring image 121 of the furrow 120 was acquired. In other words, the server 102 can identify the acquisition time of the monitoring image 121 of the furrow 120 and the acquisition location within the farm 118.

Furthermore, the data processing and transmission of the robot 101 and the data processing of the server 102 will be described in more detail with reference to FIG. 7. FIG. 7 is a diagram for explaining the relationship between the monitoring image 121 of the furrow 120 and the monitoring image 121 stored in the server 102.

When the robot 101 acquires the monitoring image 121 of the furrow 120, the robot 101 also performs a process of reading out the monitoring image 124 of the furrow 120, which is stored in the server 102 in association with the same position as the acquisition position of the monitoring image 121 of the furrow 120, to the control device 105. Note that, for ease of explanation, in the following, the monitoring image 124 of the furrow 120 read out to the control device 105 is referred to as the "read out monitoring image 124."

Then, when the robot 101 acquires the monitoring image 121 of the furrow 120, it compares the "acquired monitoring image 121 of the furrow 120" with the "read monitoring image 124." Based on the comparison result, the robot 101 determines whether or not a small obstacle (e.g., weed 122) is present.

For example, the robot 101 compares the heights of obstacles present at the same position based on the monitoring image 121 and the read monitoring image 124. If the height difference as a result of the comparison is greater than a predetermined threshold and the height of the obstacle in the monitoring image 121 is less than a predetermined first threshold, the robot 101 determines that a small obstacle (weed 122) that is a target object for weed control is present. Furthermore, if the robot 101 determines, based on the monitoring image 121 and the read monitoring image 124, that an obstacle has newly appeared in the monitoring image 121 and its height is less than the predetermined first threshold, it determines that a small obstacle is present.

If the robot 101 determines that a small obstacle is present, it continues autonomous driving. As a result, the robot 101 performs a "weed prevention" operation in which the drive wheels 114 pass over the small obstacle (weed 122) and trample the small obstacle (weed 122). The newly acquired monitoring image 121 of the furrow 120 is processed in the same manner as above and stored in the server 102.

Furthermore, the processing and transmission of data by the robot 101 and the data processing by the server 102 will be described in more detail with reference to FIG. 8. FIG. 8 is a diagram for explaining the relationship between the monitoring image 121 of the furrow 120 and the monitoring image 124 stored in the server 102.

As mentioned above, when the robot 101 acquires the monitoring image 121 of the furrow 120, the robot 101 also performs a process of reading out the monitoring image 124 of the furrow 120, which is stored in the server 102 in association with the same position as the acquisition position of the monitoring image 121 of the furrow 120, into the control device 105.

After that, when the robot 101 acquires the monitoring image 121 of the furrow 120, it compares the "acquired monitoring image 121 of the furrow 120" with the "read monitoring image 124". Based on the comparison result, the robot 101 determines whether or not a large obstacle (e.g., stone 125) is present. For example, based on the monitoring image 121 and the read monitoring image 124, the robot 101 recognizes an obstacle that is present in the monitoring image 121 but not in the read monitoring image 124. If the height of the recognized obstacle is equal to or greater than a predetermined first threshold, the robot 101 determines that a large obstacle is present.

If the robot 101 determines that a large obstacle is present, it stops moving. The newly acquired monitoring image 121 of the furrow 120 is processed in the same manner as above and stored in the server 102.

When the robot 101 stops, it uses the communication unit 109 to transmit information indicating that the robot 101 has stopped to the terminal 103 via the network N. When the terminal 103 receives the information indicating that the robot 101 has stopped, it notifies the user of the terminal 103 that the robot 101 has stopped. For example, the terminal 103 notifies the user that the robot 101 has stopped by displaying an image indicating that the robot 101 has stopped on a display screen possessed by the terminal 103. This allows the user to quickly recognize that the robot 101 has stopped due to a large obstacle, even if the user is located away from the robot 101 (for example, when not on the farm 118).

＜Effects＞
As described above, according to the agricultural support system S of the first embodiment, the robot 101 captures images of the farm 118 while autonomously traveling in the furrows 120 in the farm 118. This allows the agricultural support system S to acquire various data related to the farm 118 and the crops in the farm 118 without imposing a burden on the user (farmer) of the agricultural support system S. Furthermore, according to the agricultural support system S, if the obstacle is a small obstacle that the robot 101 can step on, the robot 101 performs a weeding operation. If the obstacle is a large obstacle that the robot 101 cannot step on, the robot 101 stops traveling to avoid a collision with the large obstacle. This allows the agricultural support system S to autonomously perform weeding while preventing the robot 101 from breaking down or interrupting weeding due to a collision with a large obstacle, thereby reducing the burden on the user's farm work (weeding work).

<<Second embodiment>>
An agricultural support system according to a second embodiment of the present invention will be described. The agricultural support system according to the second embodiment of the present invention uses a robot 200 shown in FIG. 9. The robot 200 has a configuration similar to that of the robot 101 of the first embodiment, and is characterized by being equipped with an external sensor 201 capable of detecting an obstacle as the external sensor 107. For example, a depth sensor (LiDAR (Light Detection And Ranging)), a LADAR (Laser Detection And Ranging) sensor, a ToF (Time of Flight), a stereo camera, etc. can be used for this external sensor 201. In this example, a two-dimensional LiDAR capable of scanning (scanning) laser light in the horizontal direction is used for the external sensor 201. The agricultural support system according to the second embodiment has the same configuration as the first embodiment except for the above configuration.

Figure 10 is a schematic diagram showing the robot 200 traveling on the farm 118. As shown in Figure 10, the robot 200 is required to travel along the furrows 120 of limited width without contacting or climbing over the furrows 119 while crushing the grass (not shown). Therefore, the robot 200 needs to travel accurately along a travel path 211 (a travel path 211 connecting the widthwise positions of the furrows 119 such that the robot 200 does not contact the furrows 120) so as not to contact the furrow sides 208 of the furrows 119 and obstacles 210 (see Figure 11 described below) that occur on the furrow sides 208. If the robot 200 deviates from this travel path 211, it may come into contact with the ridge side 208 or an obstacle 210, or run over the ridge 119, which may scrape the ridge side 208 and damage the ridge 119.

In response to this, the robot 200 is equipped with an external sensor 201. The robot 200 uses the external sensor 201 to detect a collision with an obstacle 210 that occurs on the ridge side 208. When the robot 200 detects the obstacle 210, it travels in a manner that avoids colliding with the obstacle 210 and prevents damage to the ridge 119.

Figure 11 is a diagram illustrating a robot 200 moving between furrows 120 while detecting furrow sides 208 and obstacles 210 using an external sensor 201.

As shown in FIG. 11, the robot 200 uses an external sensor 201 to irradiate a predetermined range in front of the robot 200 (in the direction of travel) with laser light 212 while scanning left and right in the horizontal direction. The robot 200 detects the distance and direction of the ridge side 208 and the obstacle 210 protruding from the ridge side 208 relative to the robot 200 based on the bounce time it takes for the laser light 212 to bounce back after hitting the ridge side 208 or the obstacle 210.

Furthermore, the robot 200 recognizes the shape of the ridge side 208 and the shape of the obstacle 210 in the irradiated range by running while continuously irradiating the laser light 212 from the external sensor 201. That is, when the robot 200 runs while continuously irradiating the laser light 212 from the external sensor 201, within the objects (ridge side 208, obstacle 210) in the irradiated range irradiated by the laser light 212, the bouncing time of the laser light 212 and the difference in the distance and direction of the objects relative to the robot 200 based on this difference in the distance and direction of the objects relative to the robot 200 that occurs in the irradiated range of the laser light 212.

By recognizing the distance from the furrow side 208 and the obstacle 210, the robot 200 can estimate (identify) a more precise position (relative position to the furrow 119) in the width direction of the furrow 120 in which the robot 200 exists (is traveling). For example, the robot 200 can accurately identify a more precise position of the robot 200 in the width direction of the furrow 120 by estimating (identifying) the distance from each of both sides of the furrow 120.

Therefore, the robot 200 can travel along a travel track 211 that connects the center positions of the furrows 120 in the width direction, based on the distance from each of the furrow sides 208 on both sides of the furrow 120 to the robot 200 (or the reflection time of the laser light 212).

Specifically, the robot 200 irradiates the laser light 212 onto the ridge side 208 and obstacles 210 within a predetermined range in the direction of travel, and travels while performing calculations to derive a position (i.e., travel track 211) where the reflection time from the ridge side 208 and obstacles 210 is uniform. In other words, the robot 200 is controlled by the control device 105 to travel in a position where the reflection time from the ridge side 208 is uniform, and is thereby able to travel along the travel track 211 connecting the centers of the furrows 120 in the width direction.

Furthermore, even if an obstacle 210 is present in the furrow 120, the robot 200 can travel along a travel track 211 set at the center of the width of the furrow 120 while avoiding the obstacle 210 based on the distance to the obstacle 210 and the position of the obstacle 210 detected by the external sensor 201.

That is, even if an obstacle 210 is present in the furrow 120, the robot 200 can travel along the travel path 211 set at the center of the width of the furrow 120 while avoiding the obstacle 210. Therefore, even if the furrow 120 has a meandering, curved shape, for example, the robot 200 can travel along the travel path 211 at the center of the width of the furrow 120.

Therefore, when the robot 200 travels in the furrow 120, the possibility of the robot 200 coming into contact with the furrow side 208 of the furrow 120 or with an obstacle 210 can be reduced. As a result, the robot 200 can prevent damage and partial collapse of the ridge 119 due to autonomous travel.

Figures 12A and 12B are diagrams for explaining the mounting structure of the external sensor 201. As shown in Figures 12A and 12B, when the emission direction of the laser light 212 of the external sensor 201 mounted on the robot 200 is made horizontal to the ground, the following may occur due to the fact that the height and shape of the ridges 119 differ depending on the cultivated crop.

As shown in FIG. 12A, when the height of the ridge 119 is higher than the robot 200 (more precisely, the position of the emission part 201a of the laser light 212), the laser light 212 can be irradiated onto the ridge side surface 208 of the ridge 119. Therefore, in this case, the robot 200 can recognize the ridge side surface 208 (ridge 119). In contrast, as shown in FIG. 12B, when the height of the ridge 119 is lower than the robot 200, the laser light 212 does not hit the ridge side surface 208 of the ridge 119. Therefore, in this case, the robot 200 cannot recognize the ridge side surface 208 (ridge 119).

In this way, if the emission direction of the laser light 212 of the external sensor 201 mounted on the robot 200 is made horizontal to the ground, it may happen that the robot 200 is unable to recognize the ridge side 208 of the ridge 119 due to the height and shape of the ridge 119, which change depending on the type of cultivated crop.

In contrast, the mounting structure of the external sensor 201 on the robot 200 may be movable. That is, the external sensor 201 may be mounted on the robot 200 so as to be movable with respect to the robot 200, so that the direction of the emission part 201a that emits the laser light 212 can be freely changed.

When the mounting structure of the external sensor 201 is made movable in this way, the robot 200 can freely change the emission direction of the laser light 212 of the external sensor 201 in response to changes in the height and shape of the ridge 119. This allows the robot 200 to more reliably recognize the ridge side surface 208 of the ridge 119 using the external sensor 201. Therefore, the robot 200 can recognize the ridges 119 that have various types of cultivated crops.

Figure 12C is a diagram showing an example of the configuration of the robot 200 in which the mounting structure of the external sensor 201 is movable. The external sensor 201 is attached to the robot 200 so that it can be freely tilted forward with respect to the vertical axis of the robot 200 perpendicular to the ground. That is, the external sensor 201 is attached to the robot 200 so that the axis of the external sensor 201 perpendicular to its emission part 201a (emission direction of the laser light 212) can be freely tilted forward to a predetermined angle with respect to the vertical axis of the robot 200. Figure 12C shows a state in which the external sensor 201 is attached to the robot 200 in a movable state so that it is tilted forward by about 30° with respect to the vertical axis. As shown in Figure 12C, the external sensor 201 is attached in a state in which it is tilted forward with respect to the vertical direction of the ground, so that the emission part 201a of the laser light 212 faces diagonally downward. Therefore, the robot 200 can use the external sensor 201 to irradiate the laser light 212 onto the ridge side 208 of the ridge 119 that is lower than the height of the robot 200 (more precisely, the height of the position of the emission part 201a of the laser light 212).

Figures 13 and 14 show the irradiation state of the laser light 212 when the robot 200 shown in Figure 12C travels along the furrow 120. The example shown in Figures 13 and 14 shows the irradiation state of the laser light 212 when the robot 200 travels along the furrow 120 at a constant speed in the direction indicated by the arrow a1 from time t0 and irradiates the laser light 212 from the external sensor 201 at regular time intervals (every predetermined time T1).

The external sensor 201 has its emission part 201a facing diagonally downward. Therefore, as shown in Figures 13 and 14, the laser light 212 is irradiated from the upper end to the lower end (the boundary between the ridge side surface 208 and the furrow 120) of the ridges 119 on the left and right sides of the robot 200, and across the furrow 120.

Therefore, the robot 200 can more reliably recognize (detect) the surface shape of the furrows 120, objects (obstacles) present in the furrows 120, the shape of the furrow side 208 (including obstacles 210 that occur on the furrow side 208), and the boundary between the furrows 120 and the furrow side 208 (i.e., the position of the furrow 119). Therefore, by recognizing the boundary between the furrows 120 and the furrow side 208, the robot 200 can also detect the curved furrows 120 and the obstacles 210 that have occurred in the furrows 120, and travel in the furrows 120 along a travel track 211 that connects the centers of the recognized furrows 120.

Furthermore, as described above, if the external sensor 201 is made to have a movable structure that allows it to be tilted forward, the robot 200 can more reliably recognize the furrows 119, the gaps between the furrows 120, and the obstacles 210, even when an inexpensive two-dimensional LiDAR is used as the external sensor 201.

<Robot Action 2 (Unstuck Action)>
15 is a diagram showing the state of the robot 200 stuck in the mud 222 of the furrow 120. In the furrows 120 between the furrows 119 where crops are grown, the surface of the furrows 120 can become uneven due to weather or human intervention, and water can accumulate in the unevenness, resulting in puddles or areas in a muddy, weak state (i.e., the mud 222) in the furrows 120.

If such puddles and mud 222 exist in the furrows 120 in which the robot 200 is traveling, they will impede the progress of the robot 200 (e.g., cause it to get stuck), and the robot 200 will be unable to move forward.

For example, as shown in FIG. 15, if mud 222 is present in the furrow 120 in which the robot 200 is traveling, the robot 200 may get stuck in the mud 222 and be unable to move forward (i.e., the robot 200 may become stuck).

In response to this, the robot 200 performs an operation to eliminate the stuck state (a stuck-clearing operation) as described below. FIG. 16 is a flowchart for explaining an example of a stuck-clearing operation performed by the robot 200. Each operation (process) of the robot 200 shown in FIG. 16 is executed by the control device 105.

When the robot 200 detects the halt of forward travel in step 260, the process proceeds to step 261. In step 261, the robot 200 determines whether the orientation of the robot 200 is facing the path (forward in this example). If the orientation of the robot 200 is not facing the path, the robot 200 determines "No" in step 261 and proceeds to step 262. In step 262, the robot 200 moves backward a predetermined distance and then turns to correct the orientation of the robot 200 so that it is facing the path.

If the orientation of the robot 200 is facing the direction of the path, the robot 200 judges "Yes" in step 261 and proceeds to step 263. In step 263, the robot 200 judges whether the robot 200 is not tilted (whether the robot 200 is inclined) or not.

If the robot 200 is tilted, the robot 200 determines "No" in step 263 and proceeds to step 264, where the robot 200 returns to a position where it has a normal posture (a posture that is not tilted), and then proceeds to step 265.

If the robot 200 is not tilted, the robot 200 judges "Yes" in step 263 and proceeds to step 265 to determine whether the drive wheels 114 are rotating.

If the drive wheel 114 is not rotating, the robot 200 determines "No" in step 265 and proceeds to step 266, where it stops its operation.

If the drive wheels 114 are rotating, the robot 200 judges "Yes" in step 265 and proceeds to step 267, where it attempts forward travel and judges whether there is any change in the predicted path position in the travel direction. The predicted path position is the position where the robot 200 is predicted to travel. For example, the predicted path position is a position a predetermined distance forward from the current position of the robot 200 in the travel direction of the robot 200. Therefore, if the position of the robot 200 changes (for example, it changes forward by a predetermined distance), the predicted path position also changes (for example, it changes forward by a predetermined distance).

If there is a change in the predicted path position, it is assumed that the robot 200 is not stuck or has already been resolved, so the robot 200 determines "No" in step 268 and proceeds to step 281. In step 281, the robot 200 corrects its own position (i.e., if it has deviated from the travel path 211, it corrects its position so as to return to the travel path 211). After that, the robot 200 proceeds to step 282 and starts moving forward again.

If there is no change in the predicted course position, it is possible that the robot 200 is stuck. Therefore, in this case, the robot 200 judges "Yes" in step 268 and proceeds to step 270, starts the stuck resolution process, and proceeds to step 271. The robot 200 proceeds to step 271, and performs a first operation of repeating forward and backward movement a certain number of times by rotating both of the two drive wheels 114 in the same forward rotation direction and then in the reverse rotation direction a certain number of times.

The robot 200 proceeds to step 272, where it sets the rotational speed of the drive wheels 114 to be slower than the rotational speed of the first operation, and rotates both of the two drive wheels 114 in the same forward rotation direction and in the reverse rotation direction a certain number of times, thereby performing a second operation in which forward and backward movements are repeated a certain number of times.

The robot 200 proceeds to step 273 and determines whether it is possible to move to the predicted course position. That is, the robot 200 attempts to move to the predicted course position and determines whether it was able to move to the predicted course position.

If the robot 200 cannot move to the predicted path position, the robot 200 judges "No" in step 273 and proceeds to step 274, where it starts executing a stuck-resolving turning operation in which it repeats turning at a predetermined angle. The robot 200 proceeds to step 275, where it turns in a first turning direction by a predetermined first angle (45°) by rotating one of the two driving wheels 114 in a forward rotation direction and rotating the other of the two driving wheels 114 in a backward rotation direction. Furthermore, the robot 200 turns in a second turning direction opposite to the first turning direction by a predetermined first angle (45°) by rotating one of the two driving wheels 114 in a backward rotation direction and rotating the other of the two driving wheels 114 in a forward rotation direction. Furthermore, the robot 200 turns in the first turning direction by a predetermined angle (90°), and then turns in the second turning direction by a predetermined angle (90°). The above first stuck-clearing rotation operation is repeated a specified number of times.

The robot 200 proceeds to step 276 and repeats a second unstuck turning operation a predetermined number of times, which is similar to the first unstuck turning operation, except that the rotational speed of the drive wheels 114 is changed by a predetermined speed (e.g., reduced by a predetermined speed).

When the robot 200 proceeds to step 277, it again determines whether it is possible to return (move) to the predicted course position. That is, the robot 200 attempts to move to the predicted course position and determines whether it has been able to move to the predicted course position.

If the robot 200 cannot move to the predicted course position, the robot 200 determines "No" in step 277 and proceeds to step 278, where it performs the above-mentioned first and second stack turning avoidance operations a predetermined number of times and then repeatedly determines whether the robot 200 can return to the predicted course position until a predetermined time has elapsed or until it is determined that the robot 200 can return to the predicted course position after performing the second stack turning avoidance operation a predetermined number of times. If the predetermined time has elapsed, the robot 200 stops its operation.

If it is determined that the robot 200 can return to the predicted course position, it is considered that the stuck state has been resolved, so the robot 200 returns to step 277, determines "Yes", proceeds to step 279, terminates the stuck resolution process, and proceeds to step 281. In step 281, the robot 200 corrects its own position. After that, the robot 200 proceeds to step 282, and resumes forward travel.

＜Effects＞
The agricultural support system according to the second embodiment has the same effects as the agricultural support system S according to the first embodiment. Furthermore, according to this agricultural support system, even if the robot 200 gets stuck in, for example, the mud 222, the robot 200 can autonomously resolve the stuck state. Therefore, this agricultural support system can collect data on the farm 118 without imposing a burden on the user.

<<Third embodiment>>
An agricultural support system according to a third embodiment of the present invention will now be described. The agricultural support system according to the third embodiment of the present invention uses a robot 300 shown in Fig. 17. The robot 300 has a similar configuration to the robot 101 of the first embodiment, except that the control device 105 further includes a memory device 105b and that the robot 300 includes a right ridge detection sensor 301a and a left ridge detection sensor 301b instead of the external sensor 107. The agricultural support system according to the third embodiment has a similar configuration to the first embodiment, except for the above configuration.

As shown in FIG. 18, the right ridge detection sensor 301a acquires information (e.g., the distance between the robot 300 and the object (distance information), the direction of the object) about the relative relationship between the robot 300 and objects (e.g., ridges, obstacles, etc.) present around the robot 300, and transmits the acquired information to the control device 105. The right ridge detection sensor 301a is a sensor capable of acquiring information about the relative relationship between the object and the robot 300, and is, for example, an ultrasonic sensor, a proximity sensor (magnetic sensor, etc.), an optical sensor (infrared sensor, etc.), etc.

The left ridge detection sensor 301b acquires information about the relative relationship between the robot 300 and objects (e.g., ridges, obstacles) present around the robot 300 (e.g., the distance between the robot 300 and the objects (distance information), the direction of the objects), and transmits the acquired information to the control device 105. The left ridge detection sensor 301b is a sensor capable of acquiring information about the relative relationship between the objects and the robot 300, and is, for example, an ultrasonic sensor, a proximity sensor (magnetic sensor, etc.), an optical sensor (infrared sensor, etc.), etc.

When the control device 105 acquires information from the right ridge detection sensor 301a and the left ridge detection sensor 301b, it associates it with link information (e.g., position information, object direction, object inclination, object height, etc.) related to the same object acquired by other sensors at the same time. For ease of explanation, the "information acquired from the right ridge detection sensor 301a and the left ridge detection sensor 301b" and/or the link information may also be referred to as "object information." The control device 105 controls the motor 112 based on the object information. For example, the control device 105 controls the motor 112 based on the object information so that the robot 300 runs while avoiding contact with the ridges and collisions with obstacles.

As described above, the internal sensor 108 acquires, for example, the inclination of the robot 300, the direction of movement of the robot 300, the speed of the robot 300, the geomagnetism, the direction (orientation) of the robot 300, and the rotation angle (turning angle) of the robot 300, and transmits this acquired information to the control device 105. The internal sensor 108 is, for example, a gyro sensor, an acceleration sensor, a magnetic sensor, etc.

<Robot movement 3 (moving to adjacent furrows)>
Fig. 19 is a diagram for explaining an example of the movement of the robot 300 to an adjacent furrow. As shown in Fig. 19, in the farm 118, a ridge A321a, a ridge B321b, and a ridge C321c are formed so as to be aligned in this order from left to right. A furrow A322a is formed between the adjacent ridges A321a and B321b, and a furrow B322b is formed between the adjacent ridges B321b and C321c. An example of the movement of the robot 300 from the furrow A322a to the furrow B322b in such a farm 118 will be explained below. In this example, when the robot 300 moves to the adjacent furrow B322b, the robot 300 turns 90° to the right, but it may also be configured so that when the robot 300 moves to the adjacent furrow B322b, the robot 300 turns 270° to the left.

Figure 20 is a flow chart for explaining an example of the operation of the robot 300 when it moves from furrow A 322a to the adjacent furrow B 322b. Each operation (process) of the robot 300 shown in Figure 20 is executed by the control device 105.

In step 360, the robot 300 moves straight through furrow A322a. When the robot 300 moves straight through furrow A322a (for example, when it is present at furrow movement position furrow A323a), the robot 300 detects furrow A321a (object) with the right furrow detection sensor 301a and detects furrow B321b (object) with the left furrow detection sensor 301b.

The robot 300 proceeds to step 361, where it determines whether the right ridge detection sensor 301a and the left ridge detection sensor 301b have both detected an object (ridge A 321a and ridge B 321b) while moving straight ahead.

The robot 300 passes one end of furrow A322a and moves straight ahead to a position where the right furrow detection sensor 301a and the left furrow detection sensor 301b cannot detect furrow A321a and furrow B321b. The robot 300 then enters an undetected state where neither the right furrow detection sensor 301a nor the left furrow detection sensor 301b detects an object.

Therefore, in step 361, in order to determine whether the robot 300 has crossed one end of furrow A322a, the robot 300 determines whether both the right furrow detection sensor 301a and the left furrow detection sensor 301b are in an undetected state in which no object is detected. Note that the ridge ends of ridge A321a and ridge B321b may be misaligned in the row direction due to differences in shape, collapse, etc. Therefore, the robot 300 determines whether the robot 300 has crossed one end of furrow A322a by determining whether both the right furrow detection sensor 301a and the left furrow detection sensor 301b are in an undetected state in which no object is detected.

In step 361, if at least one of the right ridge detection sensor 301a and the left ridge detection sensor 301b detects an object, the robot 300 determines "No" in step 361 and proceeds to step 362.

In step 362, the robot 300 determines whether or not the robot 300 is outside the range in which it is assumed that the ridges can be detected by at least one of the right ridge detection sensor 301a and the left ridge detection sensor 301b (hereinafter referred to as the "expected ridge range") based on the self-location information acquired by the GNSS unit 111. The robot 300 specifies the expected ridge range based on the self-location information acquired by the GNSS unit 111 and the positions of the ridges, ridge ends, and furrow spaces on the farm map stored in the ROM of the control device 105. The expected ridge range is, for example, a range consisting of the furrow space 322a and the extension range when the furrow space 322a is extended (the range of the extension part extending from the end of furrow A 322a).

If at least one of the right ridge detection sensor 301a and the left ridge detection sensor 301b detects an object even though the robot 300 is outside the expected ridge range (determination of "No" in step 361), it is considered that the robot 300 is in the state described below. That is, in this case, it is considered that there is a high possibility that the robot 300 is in a state where it is detecting an object other than the ridge (e.g., an obstacle) at a position away from the ridge ends of ridge A 321a and ridge B 321b.

Therefore, if the robot 300 is outside the expected range of the furrows, the robot 300 may not be able to move from furrow A322a to furrow B322b, so the answer is "Yes" in step 362 and the process proceeds to step 363, where an abnormality determination is made (determining that the robot 300 is in an abnormal state). If an abnormality determination is made, the robot 300 stops its operation.

In contrast, if the robot 300 is within the expected ridge range, at least one of the right ridge detection sensor 301a and the left ridge detection sensor 301b detects an object, and it is considered that the robot 300 has not completely crossed the end of the furrow 322a. Therefore, in this case, the robot 300 judges "No" in step 362 and proceeds to step 360, continuing to move straight ahead.

In step 361, if the robot 300 has completely crossed the end of the furrow 322a and both the right furrow detection sensor 301a and the left furrow detection sensor 301b no longer detect an object, the robot 300 judges "Yes" in step 361 and proceeds to step 364.

In step 364, the robot 300 moves straight a certain first specified distance stored in the storage device 105b. For convenience, this certain first specified distance may also be referred to as the "first distance". Whether the robot 300 has moved straight a certain first specified distance is determined by measuring the distance from the start position of the robot 300 moving straight (e.g., the distance obtained by integrating the speed over time) based on the internal sensor data acquired by the internal sensor 108 (the same applies to the second to fifth specified distances described later). The robot 300 may determine the first specified distance based on the self-location information and farm map acquired by the GNSS unit 111, and store the determined first specified distance in the storage device 105b (the same applies to the second to fifth specified distances described later).

If the robot 300 turns 90° to the right and then moves straight ahead at the position where both the right ridge detection sensor 301a and the left ridge detection sensor 301b no longer detect objects (ridge A321a and ridge B321b) (hereinafter referred to as the "first ridge undetected position"), there is a risk that the drive wheels 114 will run onto ridge B321b. For this reason, the robot 300 needs to turn at a position that is a first specified distance away from the first ridge undetected position in the direction of travel. Therefore, in step 364, the robot 300 moves straight ahead the first specified distance to move to turning position A323b where the robot 300 can turn without running onto ridge B321b.

Then, the robot 300 proceeds to step 365 and determines whether or not both the right ridge detection sensor 301a and the left ridge detection sensor 301b have detected an object. As a result, the robot 300 confirms, at the turning position A323b, by the right ridge detection sensor 301a and the left ridge detection sensor 301b, that there are no obstacles that would impede the movement of the robot 300 to the left and right sides.

If at least one of the right ridge detection sensor 301a and the left ridge detection sensor 301b detects an object, there is a high possibility that an obstacle that will impede the movement of the robot 300 is present, so the robot 300 judges "No" in step 365 and proceeds to step 366. In step 366, the robot 300 performs an abnormality judgment.

If neither the right ridge detection sensor 301a nor the left ridge detection sensor 301b detects an object, the robot 300 determines "Yes" in step 365 and proceeds to step 367. In step 367, the robot 300 attempts to turn 90° to the right by controlling the rotation of the drive wheel 114 at turning position A323b, and proceeds to step 368. That is, the robot 300 attempts to turn until its orientation becomes perpendicular to the row direction of the ridges.

In step 368, the robot 300 determines whether the robot 300 has been able to rotate 90° to the right based on the information obtained by the internal sensor 301c (internal sensor data).

If the robot 300 is not able to rotate 90° to the right, the robot 300 determines "No" in step 368 and proceeds to step 369, where it performs a retry within a set time previously stored in the storage device 105b. That is, in this case, the robot 300 proceeds to step 367, where it attempts to rotate until it is able to rotate 90° to the right. The time required for the retry is counted by the timer 105a. The robot 300 again proceeds to step 368, where it determines whether the robot 300 has been able to rotate 90° to the right. If the robot 300 is not able to rotate 90° to the right, it proceeds to step 369, where it determines whether a certain amount of time has elapsed since the timer 105a started counting.

If a certain period of time has not elapsed since the timer 105a started counting, the robot 300 judges "No" in step 369 and performs the operations of steps 367 and 368 again.

If a certain amount of time has elapsed since the timer 105a started counting, the robot 300 judges "Yes" and proceeds to step 370 to determine whether there is an abnormality.

If the robot 300 has been able to rotate 90° in the right direction in step 368, the robot 300 judges "Yes" in step 368 and proceeds to step 371, where it moves straight a certain second specified distance stored in the memory device 105b. That is, the robot 300 moves straight a certain second specified distance along the direction in which ridges A321a, B321b, and C321c are aligned. This certain second specified distance corresponds to the distance from the turning position A323b to the position where the ridge end of ridge B321b can be detected by the right ridge detection sensor 301a (for example, ridge end movement position A323c).

Then, the robot 300 proceeds to step 372 and determines whether the right ridge detection sensor 301a detects an object (ridge B 321b) and whether the left ridge detection sensor 301b does not detect an object. In other words, the robot 300 determines whether only the right ridge detection sensor 301a detects an object (ridge B 321b).

If the detection state of the right ridge detection sensor 301a and the left ridge detection sensor 301b is not "a state in which only the right ridge detection sensor 301a detects an object," the robot 300 judges "No" in step 372 and proceeds to step 373 to perform an abnormality judgment.

If only the right ridge detection sensor 301a detects an object, the robot 300 judges "Yes" in step 372 and proceeds to step 374, moving straight ahead a certain third specified distance stored in the memory device 105b.

The robot 300 then proceeds to step 375 and determines whether both the right ridge detection sensor 301a and the left ridge detection sensor 301b have detected an object.

If at least one of the right ridge detection sensor 301a and the left ridge detection sensor 301b detects an object, the robot 300 determines "No" in step 375 and proceeds to step 376.

When the robot 300 proceeds to step 376, it determines whether the robot 300 is present within the expected range of the furrow 322b based on the self-location information acquired by the GNSS unit 111.

The expected furrow range of furrow B322b is the extension range (the range of the extended portion extending from the end of furrow B322b) when furrow B322b is extended to include turning position B323d, and is the range in which it is expected that the right furrow detection sensor 301a and the left furrow detection sensor 301b do not detect an object (furrow) when the robot 300 is present in that range. The robot 300 identifies the expected furrow range based on its own position information acquired by the GNSS unit 111 and the positions of the furrows, furrow ends, and furrows on the farm map stored in the ROM of the control device 105.

If the robot 300 is within the expected furrow range of furrow 322b but at least one of the right furrow detection sensor 301a and the left furrow detection sensor 301b detects an object (determined as "No" in step 375), it is highly likely that there is an obstacle or the like around the robot 300 that is impeding the movement of the robot 300.

Therefore, in this case, the robot 300 judges "Yes" in step 376 and proceeds to step 377 to perform an abnormality judgment.

In contrast, if the robot 300 is outside the expected range of the furrow 322b, it is considered that the robot 300 has not reached the turning position B323d at that time. Therefore, in this case, the robot 300 judges "No" in step 376 and proceeds to step 374, and then again moves straight ahead a certain third specified distance and proceeds to step 375.

If neither the right ridge detection sensor 301a nor the left ridge detection sensor 301b detects an object in step 375, the robot 300 judges "Yes" in step 375 and proceeds to step 378, where it moves straight ahead a certain fourth specified distance stored in the memory device 105b. This certain fourth specified distance may also be referred to as the "second distance" for convenience. If the robot 300 turns 90° to the right and moves straight ahead at the position at which both the right ridge detection sensor 301a and the left ridge detection sensor 301b no longer detect an object (hereinafter referred to as the "second ridge undetected position"), there is a risk that the drive wheel 114 will run onto the ridge B321b. For this reason, the robot 300 needs to turn at a position that is the fourth specified distance away from the second ridge undetected position in the direction of travel. Therefore, in step 378, the robot 300 moves straight ahead a fourth specified distance to a turning position B323d where the robot 300 can turn without climbing onto the ridge.

Then, the robot 300 proceeds to step 379 and determines whether or not both the right ridge detection sensor 301a and the left ridge detection sensor 301b have detected an object. As a result, the robot 300 uses the right ridge detection sensor 301a and the left ridge detection sensor 301b at the turning position B323d to confirm that there are no obstacles that would hinder the movement of the robot 300 on the left and right sides.

If at least one of the right ridge detection sensor 301a and the left ridge detection sensor 301b detects an object, the robot 300 judges "No" in step 379 and proceeds to step 380 to determine whether there is an abnormality.

If neither the right ridge detection sensor 301a nor the left ridge detection sensor 301b detects an object, the robot 300 determines "Yes" in step 379 and proceeds to step 381, where it attempts to turn 90° to the right by controlling the rotation of the drive wheel 114 at turning position B323d, and proceeds to step 382. That is, the robot 300 attempts to turn until its orientation becomes parallel to the row direction of the ridges, and proceeds to step 382.

In step 382, the robot 300 determines whether the robot 300 has been able to rotate 90° to the right based on the information obtained by the internal sensor 301c (internal sensor data).

If the robot 300 is not able to rotate 90° to the right, the robot 300 judges "No" in step 382 and proceeds to step 383, where it performs a retry within a predetermined time previously stored in the storage device 105b. That is, in this case, the robot 300 proceeds to step 381, where it attempts to rotate until it is able to rotate 90° to the right. The time required for the retry is counted by the timer 105a. The robot 300 again proceeds to step 382, where it judges whether the robot 300 has been able to rotate 90° to the right. If the robot 300 is not able to rotate 90° to the right, the robot 300 judges "No" in step 382 and proceeds to step 383, where it judges whether a certain time has elapsed since the timer 105a started counting.

If a certain period of time has not elapsed since the timer 105a started counting, the robot 300 determines "No" in step 383 and performs the operations of steps 381 and 382 again.

If a certain amount of time has elapsed since the timer 105a started counting, the robot 300 judges "Yes" in step 383 and proceeds to step 384 to determine whether an abnormality has occurred.

If the robot 300 has been able to turn 90° to the right in step 382, the robot 300 determines "Yes" in step 382 and proceeds to step 385, where it moves straight ahead the fixed fifth specified distance stored in the memory device 105b. That is, the robot 300 moves straight ahead the fixed fifth specified distance toward the furrow B322b. Note that this fixed fifth specified distance corresponds, for example, to the distance from the turning position B323d to a position (for example, the furrow movement position B323e) where the right furrow detection sensor 301a and the left furrow detection sensor 301b can detect furrow B321b and furrow C321c.

The robot 300 then proceeds to step 386 and determines whether both the right ridge detection sensor 301a and the left ridge detection sensor 301b have detected an object (ridge B 321b and ridge C 321c).

If at least one of the right ridge detection sensor 301a and the left ridge detection sensor 301b does not detect an object, the robot 300 judges "No" in step 386 and proceeds to step 387 to make an abnormality judgment.

If both the right furrow detection sensor 301a and the left furrow detection sensor 301b detect an object, the robot 300 judges "Yes" in step 386 and proceeds to step 388, completing movement to the furrow 322b.

The robot 300 may also execute the operation of the flowchart of FIG. 20 described above, with step 362A described below being added between steps 361 and 364.

Step 362A: The robot 300 determines, based on the self-location information acquired by the GNSS unit 111, whether or not the robot 300 is present within the expected range of the ridge end.
The "expected ridge end range" is, for example, a range consisting of furrow space A322a and the extended range of furrow space A322a. The "expected ridge end range" is a range in which at least one of furrow ends of furrow A321a and furrow B321b can be detected by the right furrow detection sensor 301a and the left furrow detection sensor 301b when the robot 300 is present within that range. The robot 300 identifies the expected ridge end range based on its own position information acquired by the GNSS unit 111 and the positions of the furrows, furrow ends, and furrow spaces on the farm map stored in the ROM of the control device 105.

If the robot 300 is within the expected ridge end range but neither the right ridge detection sensor 301a nor the left ridge detection sensor 301b detects an object (if step 361 returns "Yes" and the process proceeds to step 362A), there may be an abnormality in the robot 300.

Therefore, in this case, the robot 300 judges "Yes" in step 362A and proceeds to step 363, determining that the robot 300 is in an abnormal state.

If the robot 300 is not within the expected range of the ridge end, the robot 300 determines "No" in step 362A and proceeds to step 364.

By executing a determination process (step 362A) based on the self-location information acquired by the GNSS unit 111 and the expected furrow end range set on the farm map, the robot 300 can more reliably determine that the robot 300 has left the furrow A322a. Furthermore, the robot 300 can determine whether or not an abnormal condition has occurred in the robot 300.

Furthermore, the robot 300 may perform an operation similar to that of the flowchart of FIG. 20 described above, in which step 376A described below is added between steps 375 and 378.

Step 376A: Based on the self-location information acquired by the GNSS unit 111, the robot 300 determines whether the robot 300 is located outside the expected furrow range of furrow B322b.

If the robot 300 is outside the expected furrow range of furrow B322b but neither the right furrow detection sensor 301a nor the left furrow detection sensor 301b detects an object (if step 375 returns "Yes" and the process proceeds to step 376A), there may be an abnormality in the robot 300.

Therefore, in this case, the robot 300 judges "Yes" in step 376A and proceeds to step 377, determining that the robot 300 is in an abnormal state.

If the robot 300 is not outside the expected furrow range of furrow B322b (i.e., the robot 322b is within the expected furrow range of furrow B322b), the robot 300 determines "No" in step 376A and proceeds to step 378.

By executing a determination process (step 376A) based on the self-location information acquired by the GNSS unit 111 and the expected furrow range set on the farm map, the robot 300 can more reliably determine that the robot 300 has entered the expected furrow range of furrow B322b. Furthermore, the robot 300 can determine whether or not an abnormal condition has occurred in the robot 300.

＜Effects＞
The agricultural support system according to the third embodiment has the same effects as the agricultural support system according to the first embodiment. Furthermore, according to this agricultural support system, the robot 300 can travel along the furrow A322a, turn at an appropriate position, and move to the adjacent furrow B322b based on the object detection results of the right furrow detection sensor 301a and the left furrow detection sensor 301b. Therefore, this agricultural support system can reduce the possibility that the robot 300, which travels autonomously within the farm 118, will come into contact with the furrow and cause a part of the furrow to collapse when the robot 300 moves to the adjacent furrow B322b.

<<Fourth embodiment>>
An agricultural support system according to a fourth embodiment of the present invention will be described. In the agricultural support system according to the fourth embodiment, as shown in FIG. 21, the robot 300 is provided with a depth sensor 301d (LiDAR (Light Detection And Ranging)) instead of the right ridge detection sensor 301a and the left ridge detection sensor 301b. The robot 300 is controlled to move between adjacent ridges based on the detection result of the object (ridge) present on the left and right sides of the robot 300 by the depth sensor 301d. Note that the robot 300 may detect the ridges present on the left and right sides of the robot 300 using a LADAR (Laser Detection And Ranging) sensor, ToF (Time of Flight), stereo camera, etc. instead of the depth sensor 301d.

<Robot movement 4 (moving to adjacent furrows)>
Fig. 22 is a diagram for explaining an example of the movement of the robot 300 to an adjacent furrow using the depth sensor 301d. Fig. 22 shows a farm 118 in which furrows A321a, B321b, and C321c similar to those in Fig. 19 are formed. Below, an example of the robot 300 moving from furrow A322a to furrow B322b in such a farm 118 will be explained. Note that in this example, the robot 300 turns right by 90° when moving to the adjacent furrow B322b, but the robot 300 may turn left by 270° when moving to the adjacent furrow B322b.

Figure 23 is a flow chart for explaining an example of the operation of the robot 300 shown in Figure 21 when it moves from furrow A 322a to the adjacent furrow B 322b. Each operation (process) of the robot 300 shown in Figure 23 is executed by the control device 105.

In step 390, the robot 300 moves straight through furrow A322a. When the robot 300 moves straight through furrow A322a (e.g., when it is at furrow movement position C333a), the robot 300 detects furrow A321a (object) on the left side of the robot 300 and furrow B321b (object) on the right side of the robot 300 using the depth sensor 301d.

The robot 300 proceeds to step 391, where it determines whether the depth sensor 301d has detected any objects (ridge A 321a and ridge B 321b) on the left and right sides of the robot 300 while moving straight ahead.

If the depth sensor 301d detects an object (at least one of ridge A 321a and ridge B 321b) present on at least one of the left and right sides of the robot 300, the robot 300 judges "No" in step 391 and proceeds to step 392.

In step 392, the robot 300 determines whether or not the robot 300 is outside the range in which the depth sensor 301d is expected to be able to detect the ridges (the "expected ridge range") based on the self-location information acquired by the GNSS unit 111. As described above, the robot 300 identifies the expected ridge range based on the self-location information acquired by the GNSS unit 111 and the positions of the ridges, ridge ends, and ridge spaces on the farm map stored in the ROM of the control device 105.

If the robot 300 is outside the expected furrow range, the robot 300 may not be able to move from furrow A322a to furrow B322b, so the robot 300 judges "Yes" in step 392 and proceeds to step 393 to judge whether there is an abnormality. If an abnormality is judged, the robot 300 stops its operation.

In contrast, if the robot 300 is not outside the expected range of the ridges, the depth sensor 301d detects an object (at least one of ridge A 321a and ridge B 321b), and it is considered that the robot 300 has not completely crossed the end of the furrow 322a. Therefore, in this case, the robot 300 judges "No" in step 392 and proceeds to step 390, continuing to move straight ahead.

In step 391, if the robot 300 has completely crossed the end of the furrow 322a and the depth sensor 301d no longer detects the object (furrow A 321a and furrow B 321b), the robot 300 judges "Yes" in step 391 and proceeds to step 394.

In step 394, the robot 300 moves straight ahead a certain first specified distance stored in the memory device 105b.

Then, the robot 300 proceeds to step 395 and determines whether the depth sensor 301d has detected any objects on the left and right sides of the robot 300. As a result, the robot 300 uses the depth sensor 301d at the turning position C333b to confirm that there are no obstacles on the left and right sides of the robot 300 that would impede its movement.

If the depth sensor 301d detects an object on at least one of the left and right sides of the robot 300, the robot 300 judges "No" in step 395 and proceeds to step 396 to determine whether there is an abnormality.

If the depth sensor 301d does not detect an object on the left or right side of the robot 300, the robot 300 judges "Yes" in step 395 and proceeds to step 397, where it attempts to turn 90° to the right at turning position C333b by controlling the rotation of the drive wheels 114, and proceeds to step 398.

In step 398, the robot 300 determines whether the robot 300 has been able to rotate 90° to the right based on the information obtained by the internal sensor 301c (internal sensor data).

If the robot 300 is not able to rotate 90° to the right, the robot 300 determines "No" in step 398 and proceeds to step 399, where it performs a retry within a set time previously stored in the storage device 105b. That is, in this case, the robot 300 proceeds to step 397, where it continues to try to rotate until it is able to rotate 90° to the right. The time required for the retry is counted by the timer 105a. The robot 300 again proceeds to step 398, where it determines whether the robot 300 has been able to rotate 90° to the right. If the robot 300 is not able to rotate 90° to the right, it proceeds to step 399, where it determines whether a certain amount of time has elapsed since the timer 105a started counting.

If a certain period of time has not elapsed since the timer 105a started counting, the robot 300 judges it as "No" and performs the operations of steps 397 and 398 again.

If a certain amount of time has elapsed since the timer 105a started counting, the robot 300 judges "Yes" and proceeds to step 400 to determine whether there is an abnormality.

If the robot 300 has been able to turn 90° to the right in step 398, the robot 300 determines "Yes" in step 398 and proceeds to step 401, where it moves straight ahead the fixed second specified distance stored in the memory device 105b. Note that this fixed second specified distance corresponds to the distance from the turning position C333b to the position where the ridge end of the ridge B321b can be detected by the depth sensor 301d (e.g., the ridge end movement position B333c).

The robot 300 then proceeds to step 402 and determines whether the depth sensor 301d detects only objects present on the right side of the robot 300.

If the detection state of the depth sensor 301d is not a state in which the depth sensor 301d is detecting only objects present on the right side of the robot 300, the robot 300 judges "No" in step 402 and proceeds to step 403 to judge an abnormality.

If the depth sensor 301d detects only objects present to the right of the robot 300, the robot 300 judges "Yes" in step 402 and proceeds to step 404, where it moves straight ahead a certain third specified distance stored in the memory device 105b.

The robot 300 then proceeds to step 405 and determines whether the depth sensor 301 has detected any objects to the left and right sides of the robot 300.

If the depth sensor 301d detects an object on at least one of the left and right sides of the robot 300, the robot 300 determines "No" in step 405 and proceeds to step 406.

When the robot 300 proceeds to step 406, it determines whether or not the robot 300 is within the expected furrow range of the furrow 322b based on the self-location information acquired by the GNSS unit 111. As described above, the robot 300 identifies the expected furrow range based on the self-location information acquired by the GNSS unit 111 and the positions of the furrows, furrow ends, and furrows on the farm map stored in the ROM of the control device 105.

If the depth sensor 301d detects an object even though the robot 300 is within the expected range of the furrow 322b (determination of "No" in step 405), it is highly likely that there is an obstacle or the like around the robot 300 that is impeding the movement of the robot 300. Therefore, in this case, the robot 300 determines "Yes" in step 406 and proceeds to step 407 to perform an abnormality determination.

In contrast, if the robot 300 is outside the expected range of the furrow 322b, the robot 300 determines "No" in step 406 and proceeds to step 404, then moves straight ahead again the third specified distance and proceeds to step 405.

If the depth sensor 301d does not detect an object on the left or right side of the robot 300 in step 405, the robot 300 judges "Yes" in step 405 and proceeds to step 408, where it moves straight a certain fourth specified distance stored in the memory device 105b. If the robot 300 turns and moves straight at the position at which the depth sensor 301d no longer detects an object (the "second ridge undetected position"), there is a risk that the drive wheel 114 will run onto the ridge B321b. For this reason, the robot 300 needs to turn at a position that is the fourth specified distance away in the direction of travel from the second ridge undetected position. Therefore, in step 408, the robot 300 moves straight a fourth specified distance to a turning position D333d where the robot 300 can turn without running onto the ridge.

Then, the robot 300 proceeds to step 409 and determines whether the depth sensor 301d has detected an object. As a result, the robot 300 uses the depth sensor 301d at the turning position D333d to confirm that there are no obstacles on the left and right sides of the robot 300 that would impede the movement of the robot 300.

If the depth sensor 301d detects an object on at least one of the left and right sides of the robot 300, the robot 300 judges "No" in step 409 and proceeds to step 410 to determine whether there is an abnormality.

If the depth sensor 301d does not detect an object on the left or right side of the robot 300, the robot 300 judges "Yes" in step 409 and proceeds to step 411, where it attempts to turn 90° to the right at turning position D333d by controlling the rotation of the drive wheels 114, and proceeds to step 412.

In step 412, the robot 300 determines whether the robot 300 has been able to rotate 90° to the right based on the information acquired by the internal sensor 301c (internal sensor data).

If the robot 300 is unable to rotate 90° to the right, the robot 300 determines "No" in step 412 and proceeds to step 413, where it performs a retry within a set time previously stored in the storage device 105b. That is, in this case, the robot 300 proceeds to step 411, where it attempts to rotate until it is able to rotate 90° to the right. The time required for the retry is counted by the timer 105a. The robot 300 again proceeds to step 412, where it determines whether the robot 300 has been able to rotate 90° to the right. If the robot 300 is unable to rotate 90° to the right, the robot 300 determines "No" in step 412 and proceeds to step 413, where it determines whether a certain time has elapsed since the timer 105a started counting.

If a certain period of time has not elapsed since the timer 105a started counting, the robot 300 determines "No" in step 413 and performs the operations of steps 411 and 412 again.

If a certain amount of time has elapsed since the timer 105a started counting, the robot 300 judges "Yes" in step 413 and proceeds to step 414 to determine whether there is an abnormality.

If the robot 300 has been able to turn 90° to the right in step 412, the robot 300 determines "Yes" in step 412 and proceeds to step 415, where it moves straight ahead a fixed fifth specified distance stored in the memory device 105b. Note that this fixed fifth specified distance corresponds, for example, to the distance from the turning position D333d to a position where the depth sensor 301d can detect ridge B321b and ridge C321c (for example, the inter-ridge movement position D333e).

The robot 300 then proceeds to step 416 and determines whether the depth sensor 301d detects objects (ridge B 321b and ridge C 321c) on the left and right sides of the robot 300.

If the depth sensor 301d does not detect an object on at least one of the left and right sides of the robot 300, the robot 300 judges "No" in step 416 and proceeds to step 417 to determine whether an abnormality has occurred.

If the depth sensor 301d detects objects (furrow B 321b and furrow C 321c) on the left and right sides of the robot 300, the robot 300 judges "Yes" in step 416 and proceeds to step 418, completing movement to the furrow 322b.

The robot 300 may also execute an operation similar to that of the flowchart in FIG. 23, except that step 392A described below is added between steps 391 and 394.

Step 392A: The robot 300 determines, based on the self-location information acquired by the GNSS unit 111, whether or not the robot 300 is present within the expected range of the ridge end.
The "expected ridge end range" is a range in which at least one of ridge A 321a and ridge B 321b is expected to be detectable by the depth sensor 301d when the robot 300 is present within that range. As described above, the robot 300 specifies the expected ridge end range based on its own position information acquired by the GNSS unit 111 and the positions of the ridges, ridge ends, and spaces between ridges on the farm map stored in the ROM of the control device 105.

If the depth sensor 301d does not detect objects (ridge A 321a and ridge B 321b) on the left and right sides of the robot 300 even though the robot 300 is within the expected range of the ridge end (if the answer to step 391 is "Yes" and the process proceeds to step 392A), there may be an abnormality in the robot 300.

Therefore, in this case, the robot 300 judges "Yes" in step 392A and proceeds to step 393, determining that the robot 300 is in an abnormal state.

If the robot 300 is not within the expected range of the ridge end, the robot 300 determines "No" in step 392A and proceeds to step 394.

By executing a determination process (step 392A) based on the self-location information acquired by the GNSS unit 111 and the expected furrow end range set on the farm map, the robot 300 can more reliably determine that the robot 300 has left the furrow A322a. Furthermore, the robot 300 can determine whether or not an abnormal condition has occurred in the robot 300.

Furthermore, the robot 300 may be configured to execute an operation similar to that of the flowchart of FIG. 23 described above, in which step 406A described below is added between step 405 and step 408.

Step 406A: Based on the self-location information acquired by the GNSS unit 111, the robot 300 determines whether the robot 300 is located outside the expected furrow range of furrow B322b.

If the robot 300 is outside the expected furrow range of furrow B322b but the depth sensor 301d does not detect any objects on the left or right side of the robot 300 (if the answer in step 405 is "Yes" and the process proceeds to step 406A), there may be an abnormality in the robot 300.

Therefore, in this case, the robot 300 judges "Yes" in step 406A and proceeds to step 407, determining that the robot 300 is in an abnormal state.

If the robot 300 is not outside the expected range of furrow B322b (i.e., the robot 300 is within the expected range of furrow B322b), the robot 300 determines "Yes" in step 406A and proceeds to step 408.

By executing a determination process (step 406A) based on the self-location information acquired by the GNSS unit 111 and the expected furrow range set on the farm map, the robot 300 can more reliably determine that the robot 300 has entered the expected furrow range of furrow B322b. Furthermore, the robot 300 can determine whether or not an abnormal condition has occurred in the robot 300.

＜Effects＞
The agricultural support system according to the fourth embodiment has the same effects as the agricultural support system according to the third embodiment.

<<Fifth embodiment>>
An agricultural support system according to a fifth embodiment of the present invention will be described. As shown in Fig. 24, the agricultural support system according to the fifth embodiment of the present invention has markers A351a and B351b, which indicate turning positions, installed at predetermined positions in a farm 118. A robot 300 shown in Fig. 21 recognizes the markers A351a and B351b by a monitoring device 106 and a depth sensor 301d, and thereby performs a movement operation from the furrow A322a to the furrow B322b.

In farm 118, marker A351a is placed on the extension of furrow A322a. Marker B351b is placed on the extension of furrow C321c.

The robot 300 measures the distance from the robot 300 to the marker A351a using the depth sensor 301d while moving straight through the furrow 322a. The robot 300 turns 90° to the right at turning position E343b when the measured distance becomes a certain distance set in the storage device 105b.

Then, while moving straight, the robot 300 measures the distance from the robot 300 to the marker B351b using the depth sensor 301d. The robot 300 turns 90° to the right at turning position E343d where the measured distance becomes a fixed distance set in the memory device 105b. The robot 300 then moves straight to enter furrow B322b, completing the movement from furrow A322a to furrow B322b.

Note that marker A351a and marker B351b can have unique information. For example, information indicating the installation location of the marker can be written in characters or a QR code (registered trademark), and the robot 300 can obtain its own accurate position by reading it with the monitoring device 106. Furthermore, if the robot 300 determines that an error has occurred in the self-position information obtained by the GNSS unit 311 based on the read information, it becomes possible to correct the self-position information.

<Robot movement 5 (moving to adjacent furrows)>
25 is a flowchart for explaining an example of the movement of the robot 300 to an adjacent space between furrows in the farm 118 in which a marker is installed. Each operation (process) of the robot 300 is executed by the control device 105.

In step 420, the robot 300 moves straight through the furrow A322a. The robot 300 proceeds to step 421, recognizes the marker A351a using the monitoring device 106, and moves to the front of the marker A351a. The robot 300 measures the distance to the marker A351a using the depth sensor 301d, and determines whether the robot 300 is at the turning position E343b based on the distance to the marker A351a and the distance set in the memory device 105b.

If the robot 300 is not at the turning position E343b, the robot 300 determines "No" in step 421 and proceeds to step 422, where it determines whether the robot 300 is outside the expected range of the turning position E343b based on the self-location information acquired by the GNSS unit 311. Note that the expected range of the turning position E343b is the range within which the robot 300 can recognize the marker A351a using the depth sensor 301d when the robot 300 is present within that range. For convenience, the "expected range of the turning position E343b" may also be referred to as the "expected range for first marker detection."

If the robot 300 is within the expected range of the turning position E343b, the robot 300 judges "No" in step 422 and proceeds to step 420, where the robot 300 again moves forward in a straight line and attempts to move to the turning position E343b.

If the robot 300 is outside the expected range of the rotation position E343b, the robot 300 judges "Yes" in step 422 and proceeds to step 423 to judge an abnormality. That is, even though the robot 300 is outside the expected range in which the robot 300 can recognize the marker A351a, the robot 300 is able to detect the marker A351a (judgment "No" in step 421), so the robot 300 judges an abnormality.

If the robot 300 is in the turning position E343b, the robot 300 judges "Yes" in step 421 and proceeds to step 424, where it attempts to turn 90° to the right by controlling the rotation of the drive wheels 114 at the turning position E343b, and proceeds to step 425.

In step 425, the robot 300 determines whether the robot 300 has been able to rotate 90° to the right based on the information acquired by the internal sensor 301c (internal sensor data).

If the robot 300 is not able to rotate 90° to the right, the robot 300 determines "No" in step 425 and proceeds to step 426, where it performs a retry within a set time previously stored in the storage device 105b. That is, in this case, the robot 300 proceeds to step 424, where it attempts to rotate until it is able to rotate 90° to the right. The time required for the retry is counted by the timer 105a. The robot 300 again proceeds to step 425, where it determines whether the robot 300 has been able to rotate 90° to the right. If the robot 300 is not able to rotate 90° to the right, it proceeds to step 426, where it determines whether a certain amount of time has elapsed since the timer 105a started counting.

If a certain period of time has not elapsed since the timer 105a started counting, the robot 300 judges it as "No" and performs the operations of steps 424 and 425 again.

If a certain amount of time has elapsed since the timer 105a started counting, the robot 300 judges "Yes" in step 426 and proceeds to step 427 to perform an abnormality judgment.

If the robot 300 is able to turn 90° to the right in step 425, the robot 300 judges "Yes" in step 425 and proceeds to step 428, moving straight toward marker B351b.

The robot 300 proceeds to step 429, recognizes marker B351b using the monitoring device 106, and moves to the front of marker B351b. The robot 300 measures the distance to marker B351b using the depth sensor 301d, and determines whether the robot 300 is at the turning position F343d based on the distance to marker B351b and the distance set in the storage device 105b.

If the robot 300 is not at the turning position F343d, the robot 300 determines "No" in step 429 and proceeds to step 430, where it determines whether the robot 300 is outside the expected range of the turning position F343d based on the self-location information acquired by the GNSS unit 311. Note that the expected range of the turning position F343d is the range in which the robot 300 can recognize the marker B351b using the depth sensor 301d if the robot is present within that range. For convenience, the expected range of the turning position F343d may also be referred to as the "second marker detection expected range."

If the robot 300 is within the expected range of the turning position F343d, the robot 300 determines "No" in step 430 and proceeds to step 428, where the robot 300 again moves forward in a straight line and attempts to move to the turning position F343d.

If the robot 300 is outside the expected range of the rotation position F343d, the robot 300 judges "Yes" in step 430 and proceeds to step 431 to judge an abnormality. That is, since the robot 300 is able to detect the marker B351b even though it is outside the expected range in which the robot 300 can recognize the marker B351b (step 429), the robot 300 judges an abnormality.

If the robot 300 is in the turning position F343d, the robot 300 judges "Yes" in step 429 and proceeds to step 432, where it attempts to turn 90° to the right by controlling the rotation of the drive wheels 114 at the turning position F343d, and proceeds to step 433.

In step 433, the robot 300 determines whether the robot 300 has been able to rotate 90° to the right based on the information acquired by the internal sensor 301c (internal sensor data).

If the robot 300 is not able to rotate 90° to the right, the robot 300 determines "No" in step 433 and proceeds to step 434, where it performs a retry within a set time previously stored in the storage device 105b. That is, in this case, the robot 300 proceeds to step 432, where it attempts to rotate until it is able to rotate 90° to the right. The time required for the retry is counted by the timer 105a. The robot 300 again proceeds to step 433, where it determines whether the robot 300 has been able to rotate 90° to the right. If the robot 300 is not able to rotate 90° to the right, it proceeds to step 434, where it determines whether a certain amount of time has elapsed since the timer 105a started counting.

If a certain period of time has not elapsed since the timer 105a started counting, the robot 300 judges it as "No" and performs the operations of steps 432 and 433 again.

If a certain amount of time has elapsed since the timer 105a started counting, the robot 300 judges "Yes" in step 434 and proceeds to step 435 to determine whether there is an abnormality.

If the robot 300 has been able to rotate 90° to the right in step 433, the robot 300 judges "Yes" in step 433 and proceeds to step 446, where it moves straight ahead and completes its movement to the furrow 322b.

＜Effects＞
The agricultural support system according to the fifth embodiment has the same effects as the agricultural support system according to the first embodiment. Furthermore, according to this agricultural support system, the robot 300 can move to the adjacent furrow B322b by traveling through the furrow A322a and then turning at an appropriate position based on the distance between the robot 300 and the markers (marker A351a, marker B351b). Therefore, this agricultural support system can reduce the possibility that the robot 300, which travels autonomously within the farm 118, will come into contact with the furrow and cause a part of the furrow to collapse when the robot 300 moves to the adjacent furrow B322b.

<<Modifications>>
The present invention is not limited to the above-described embodiments, and various modifications can be adopted within the scope of the present invention. Furthermore, the features of the above-described embodiments can be combined with each other without departing from the scope of the present invention.

For example, in each of the above-described embodiments, the robot may be a moving body or crawler equipped with multiple pairs of drive wheels. When multiple robots exist in a farm, one robot may use information including the self-location information of the other robots.

101...robot, 102...server, 103...terminal, 105...control device, 106...monitoring device, 107...external sensor, 108...internal sensor, 109...communication unit, 110...antenna, 111...GNSS unit, 112...motor, 113...drive shaft, 114...drive wheel, 118...farm, 119...furrow, 120...furrow spacing, 121...monitoring image, 123...farm map data, 301a...right furrow detection sensor, 301b...left furrow detection sensor, 322a...furrow spacing A, 322b...furrow spacing B, S...agricultural support system

Claims (15)

An agricultural support system including an information processing device and a mobile body capable of traveling on a farm formed with a plurality of rows of ridges and furrow spaces, the information processing device and the mobile body being configured to be able to communicate with each other,
The moving body is
A communication device capable of communicating with the information processing device;
A memory unit in which a farm map corresponding to the farm is stored, the farm map including ridge information including positions of the plurality of ridges and the ridge spaces;
a first information acquisition device that detects objects by acquiring object information regarding objects around the moving object;
A second information acquisition device that acquires moving object information related to a state of the moving object;
a position information acquisition device that acquires self-position information indicating a position of the moving object;
A traveling device capable of causing the moving body to travel in a desired traveling direction;
a control device that controls the traveling device so that the moving body autonomously travels between the furrows based on the farm map, the object information, the moving body information, and the self-location information;
Equipped with
The control device includes:
When the moving body is traveling between one of the furrows, if the first information acquisition device changes from a ridge detection state in which the moving body detects the ridge on the left and right sides to a ridge non-detection state in which the moving body does not detect the ridge on the left and right sides,
controlling the traveling device so that the moving body moves in a straight line a first distance from a position at which the first information acquisition device entered the ridge undetected state to a first turning position, turns at the first turning position, and then performs a first straight line movement in which the moving body moves in a straight line along a direction in which the multiple ridges are arranged toward a position where the ridges are present only on one of the left and right sides of the moving body;
Furthermore, the control device
When the first information acquisition device goes from a state in which the ridge is detected only on the one side of the moving body to a ridge undetected state in which the ridge is not detected on the left and right sides of the moving body,
the movable body moves in a straight line a second distance from the position at which the furrow undetected state occurred to a second turning position, turns at the second turning position, and then performs a second straight movement operation of moving in a straight line toward the furrow adjacent to one of the furrows,
Controlling the traveling device
It was configured as follows:
Agricultural support system. An agricultural support system including an information processing device and a mobile body capable of traveling on a farm formed with a plurality of rows of ridges and furrow spaces, the information processing device and the mobile body being configured to be able to communicate with each other,
The moving body is
A communication device capable of communicating with the information processing device;
A memory unit in which a farm map corresponding to the farm is stored, the farm map including ridge information including positions of the plurality of ridges and the ridge spaces;
A first information acquisition device for acquiring object information regarding objects around the moving object;
A second information acquisition device for acquiring moving body information relating to a state of the moving body;
a position information acquisition device that acquires self-position information indicating a position of the moving object;
A traveling device capable of causing the moving body to travel in a desired traveling direction;
a control device that controls the traveling device so that the moving body autonomously travels between the furrows based on the farm map, the object information, the moving body information, and the self-location information;
Equipped with
A first marker is installed at a position separated by a first predetermined distance from a first turning position, and a second marker is installed at a position separated by a second predetermined distance from a second turning position,
The control device includes:
When the moving body is traveling between one of the furrows, the first information acquisition device acquires a first marker distance, which is a distance between the first marker and the moving body, as the object information;
When it is determined that the moving object has reached the first turning position based on the first marker distance and the first predetermined distance,
controlling the traveling device so that, after the moving body has turned at the first turning position, the moving body moves straight along a direction in which the plurality of ridges are arranged toward a position where the ridges are present only on one of the left side and the right side of the moving body;
The first information acquisition device acquires a second marker distance, which is a distance between the second marker and the moving body, as the object information;
when it is determined that the moving body has reached the second turning position based on the second marker distance and the second predetermined distance, the traveling device is controlled so that the moving body turns at the second turning position and then moves straight toward the furrow adjacent to the one of the furrows.
It was configured as follows:
Agricultural support system. In the agricultural support system according to claim 1 or 2,
The information processing device further includes a storage device in which the farm map is stored,
The moving body further includes an imaging device that captures an image of the periphery of the moving body to obtain a captured image,
The control device includes:
receiving the captured image from the imaging device;
a data set in which the captured image, time information indicating a time when the captured image was acquired, and image acquisition position information indicating a position when the captured image was acquired are associated with each other is transmitted to the information processing device via the communication device;
The information processing device includes:
based on the image acquisition position information, the captured image and the time information are stored in the storage device in association with the farm map;
It was configured as follows:
Agricultural support system. In the agricultural support system according to claim 3,
The control device includes:
When the captured image is received from the imaging device, a saved image is acquired from the information processing device via the communication device, the saved image being the captured image stored in the storage device of the information processing device and corresponding to a position where the captured image was acquired;
By comparing the captured image with the stored image, it is determined whether an object present in the captured image is a weed control target object or an obstacle;
Controlling the travel of the moving object based on the determination result.
It was configured as follows:
Agricultural support system. In the agricultural support system according to claim 4,
The control device includes:
If the object present in the captured image is the weed control target object, the moving body continues traveling;
When the object present in the captured image is the obstacle, the traveling device is controlled to stop the traveling of the moving body.
It was configured as follows:
Agricultural support system. In the agricultural support system according to claim 1 or 2,
The control device includes:
When it is determined that the moving body is in a stuck state based on the moving body information and the self-position information, the traveling device is controlled so as to cause the moving body to perform a predetermined unstuck operation for unstuck the moving body.
It was configured as follows:
Agricultural support system. In the agricultural support system according to claim 1 or 2,
the moving body includes a sensor having an emission unit that emits laser light as the first information acquisition device,
The sensor is attached to the movable body so as to be movable relative to the movable body so that the direction of the emission portion can be freely changed.
Agricultural support system. In the agricultural support system according to claim 1 or 2,
The information processing device further includes an information terminal capable of communicating with the information processing device and the mobile object,
The information terminal acquires information indicating a state of the mobile object from the mobile object.
It was configured as follows:
Agricultural support system. In the agricultural support system according to claim 1 or 2,
The control device includes:
controlling the traveling device so that the moving body turns at the first turning position until the orientation of the moving body becomes perpendicular to the row direction of the ridges;
controlling the traveling device so that the moving body turns at the second turning position until the orientation of the moving body becomes parallel to the row direction of the ridges;
It was configured as follows:
Agricultural support system. In the agricultural support system according to claim 1 ,
The control device includes:
determining whether or not the first information acquisition device detects an object on the left and right sides of the moving body before the moving body turns at the first turning position;
When the first information acquisition device does not detect an object on the left or right side of the moving body, the traveling device is controlled so that the moving body performs the first straight forward motion;
When the first information acquisition device detects an object on at least one of a left side and a right side of the moving body, the traveling device is controlled to stop the moving body;
determining whether or not the first information acquisition device detects an object on the left and right sides of the moving body before the moving body turns at the second turning position;
When the first information acquisition device does not detect an object on the left or right side of the moving body, the traveling device is controlled so that the moving body performs the second straight forward motion;
When the first information acquisition device detects an object on at least one of the left and right sides of the moving body, the traveling device is controlled to stop the moving body.
It was configured as follows:
Agricultural support system. In the agricultural support system according to claim 1 ,
The control device, based on the self-location information and the farm map, identifies an expected ridge range consisting of one of the furrows and an extension range of the one of the furrows when the one of the furrows is extended, and when the moving body is present within the expected ridge range, it is assumed that at least one of the two furrows sandwiching the one of the furrows can be detected by the first information acquisition device;
determining whether the moving object is located outside the assumed ridge range based on the self-location information;
When the first information acquisition device detects an object other than the ridge on at least one of the left and right sides of the moving body, if the moving body is outside the expected ridge range, the moving body is determined to be in an abnormal state.
It was configured as follows:
Agricultural support system. In the agricultural support system according to claim 1 ,
The control device includes:
based on the self-location information and the farm map, a ridge end assumed range consisting of one of the furrow spaces and an extension range of the one of the furrow spaces when the one of the furrow spaces is extended, and when the moving body is present within the ridge end assumed range, the first information acquisition device identifies the ridge end assumed range in which it is assumed that the ridge end of at least one of the two ridges sandwiching the one of the furrow spaces can be detected;
when the first information acquisition device changes from the ridge detection state to the ridge non-detection state while the moving body is traveling between the one of the furrows, it determines whether or not the moving body is present within the ridge end expected range based on the self-position information;
When the moving object is present within the ridge end assumed range, it is determined that the moving object is in an abnormal state.
It was configured as follows:
Agricultural support system. In the agricultural support system according to claim 1 ,
The control device includes:
based on the self-location information and the farm map, a furrow expected range is identified, which is an extension range of the furrows adjacent to one of the furrows, and in which it is expected that the furrows will not be detectable by the first information acquisition device if the moving body is present within the extension range;
When the first information acquisition device detects an object other than the ridge on at least one of the left and right sides of the moving body, if the moving body is present within the expected furrow spacing range, the moving body is determined to be in an abnormal state.
It was configured as follows:
Agricultural support system. In the agricultural support system according to claim 2,
The control device includes:
determining a first marker detection assumed range in which the first marker is assumed to be detectable by the first information acquisition device and a second marker detection assumed range in which the second marker is assumed to be detectable by the first information acquisition device based on the self-location information and the farm map;
When the first information acquisition device has acquired the first marker distance, before it is determined that the moving object has reached the first turning position, it is determined whether or not the moving object is located outside an assumed detection range of the first marker based on the self-location information;
When the moving object is present outside the first marker detection assumed range, the moving object is determined to be in an abnormal state;
When the second marker distance can be acquired by the first information acquisition device, before it is determined that the moving object has reached the second turning position, it is determined whether or not the moving object is present outside an assumed detection range of the second marker based on the self-location information;
When the moving object is present outside the second marker detection assumed range, the moving object is determined to be in an abnormal state.
It was configured as follows:
Agricultural support system. A mobile body configured to be capable of communicating with an external information processing device and capable of traveling on a farm formed with a plurality of rows of ridges and furrow spaces,
A communication device capable of communicating with the information processing device;
A memory unit in which a farm map corresponding to the farm is stored, the farm map including ridge information including positions of the plurality of ridges and the ridge spaces;
a first information acquisition device that detects objects by acquiring object information regarding objects around the moving object;
A second information acquisition device that acquires moving object information related to a state of the moving object;
a position information acquisition device that acquires self-position information indicating a position of the moving object;
A traveling device capable of causing the moving body to travel in a desired traveling direction;
a control device that controls the traveling device so that the moving body autonomously travels between the furrows based on the farm map, the object information, the moving body information, and the self-location information;
Equipped with
The control device includes:
When the moving body is traveling between one of the furrows, if the first information acquisition device changes from a ridge detection state in which the moving body detects the ridge on the left and right sides to a ridge non-detection state in which the moving body does not detect the ridge on the left and right sides,
controlling the traveling device so that the moving body moves in a straight line a first distance from a position at which the first information acquisition device entered the ridge undetected state to a first turning position, turns at the first turning position, and then performs a first straight line movement in which the moving body moves in a straight line along a direction in which the multiple ridges are arranged toward a position where the ridges are present only on one of the left and right sides of the moving body;
Furthermore, the control device
When the first information acquisition device goes from a state in which the ridge is detected only on the one side of the moving body to a ridge undetected state in which the ridge is not detected on the left and right sides of the moving body,
the movable body moves in a straight line a second distance from the position at which the furrow undetected state occurred to a second turning position, turns at the second turning position, and then performs a second straight movement operation of moving in a straight line toward the furrow adjacent to one of the furrows,
Controlling the traveling device
It was configured as follows:
Mobile body.

Images