JP7554726B2 - Method for moving and searching for unmanned vehicle - Google Patents

Description

The present invention relates to a method for moving and exploring an unmanned vehicle.

A conventional technique in this field is the inspection system described in Patent Document 1 below. In this inspection system, an unmanned airship equipped with inspection equipment is flown in the internal space of a boiler furnace. The unmanned airship then moves to each part to be inspected inside the boiler furnace, and the inspection equipment acquires external images, thermal images, sound, wind speed, and the like. In this inspection system, the current position of the unmanned airship is recognized using GPS technology. Alternatively, the unmanned airship receives a signal from a transmitter installed, for example, at the four corners of the boiler furnace, and the current position of the unmanned airship is recognized by three-dimensional position measurement based on this signal, thereby controlling the flight of the unmanned airship.

JP 2004-211995 A

However, in exploration using this type of unmanned mobile body, the target space may be a place where GPS satellite signals cannot reach, and there are cases where the Global Navigation Satellite System (GNSS) cannot be used. Furthermore, in this type of exploration, the target space may be a space where the environment cannot be known in advance. In such cases, the safety of the target space has not been confirmed, so people cannot enter in advance and it is difficult to install equipment in the target space in advance. As a result, it is not possible to control the movement of the unmanned mobile body by relying on equipment in the target space.

In view of these problems, the present invention aims to provide a movement method and exploration method that enables an unmanned moving body to move within a target area without requiring a GNSS signal or any equipment to be installed in the target space in advance.

The method for moving an unmanned mobile body of the present invention is a method for moving an unmanned mobile body through a specified target space that spreads forward in the direction of movement from a specified starting point, and includes a first step, a second step, a third step, ... in which a portion of the area in the target space is mapped, and in the kth step (where k = 2, 3, ...), the unmanned mobile body moves from a state in which it is installed at the starting point to the (k-1)th waypoint stored in the (k-1)th step, maps the kth area in front of the (k-1)th waypoint by SLAM processing, and returns to the starting point, and the kth waypoint set within the mapped kth area is stored in the unmanned mobile body.

In the method for moving an unmanned mobile body of the present invention, at least three physical targets detectable by the SLAM processing of the unmanned mobile body are fixed near the starting point, and in the kth step (where k = 1, 2, ...), the unmanned mobile body may recognize its own position based on the positions of the physical targets detected by the SLAM processing near the starting point.

In addition, in the kth step (where k = 1, 2, ...), a movement simulation of the unmanned mobile body is performed based on the mapping information of the kth area stored by the unmanned mobile body that has returned to the starting point, and a point at which the movement simulation determines that the unmanned mobile body can move back and forth from the starting point is set as the kth waypoint.

In addition, in the kth step (where k=3, 4, ...), the unmanned moving body may move from the starting point to the (k-1)th waypoint along a movement path that passes through the first to (k-1)th waypoints in order from the starting point. The unmanned moving body may be an unmanned aerial vehicle capable of flying within a target space and equipped with a scanning device for SLAM processing. The unmanned aerial vehicle may be equipped with a main body having a power source, a pair of landing legs extending downward from the main body, and a beam section that spans between the landing legs below the main body, and the scanning device is installed on the beam section.

In the exploration method of the present invention, an unmanned mobile body equipped with sensors that acquire environmental information within a target space moves within the target space using any of the above-mentioned methods for moving an unmanned mobile body, and environmental information obtained by the sensors while the unmanned mobile body is moving within the target space is collected.

The present invention provides a movement method and exploration method that enables an unmanned moving body to move within a target area without requiring a GNSS signal or any equipment to be installed in the target space in advance.

1 is a cross-sectional view showing a shield tunnel to which the unmanned mobile body movement method and exploration method of the present embodiment are applied. 1A is a block diagram of an unmanned exploration system, and FIG. 1B is a block diagram showing the physical configuration of an overall control unit of an unmanned aerial vehicle. 1A is a perspective view showing the vicinity of the starting point, and FIG. 1B is an enlarged perspective view showing the target. FIG. 2 is a front view of the unmanned aerial vehicle. 4 is a flowchart showing a method for detecting a shield tunnel according to the present embodiment. Following FIG. 5, there is shown a flowchart illustrating the method for detecting a shield tunnel according to the present embodiment.

Below, an embodiment of the movement method and exploration method for an unmanned mobile body according to the present invention will be described in detail with reference to the drawings. FIG. 1 is a vertical cross-sectional view showing an example of a target space of the movement method and exploration method according to the present invention. In this embodiment, a shield tunnel 101 formed underground by a shield machine will be described as an example of the target space. The shield tunnel 101 extends in a generally horizontal direction from the bottom of a vertical shaft 103. Within this shield tunnel 101, there may be, for example, areas that may collapse, areas that are submerged, areas where rubble is scattered, areas where toxic gases are present, etc., and the safety of the inside of the shield tunnel 101 has not been confirmed.

In order to check the environment inside the shield tunnel 101 in an unmanned manner, an unmanned exploration system 1 equipped with an unmanned aerial vehicle 3 (unmanned mobile vehicle, unmanned aerial vehicle) is used as shown in FIG. 2(a). Since the shield tunnel 101 is an underground space, flight control of the unmanned aerial vehicle 3 using GNSS signals is not possible. In addition, since people cannot enter the shield tunnel 101 in advance, it is not possible to install equipment, etc. in the shield tunnel 101 in advance. Therefore, it is not possible to control the flight of the unmanned aerial vehicle 3 using equipment, etc. (for example, signs that the unmanned aerial vehicle 3 can recognize, relay points for wireless communication, etc.) already installed in the shield tunnel 101. In addition, since the shield tunnel 101 is an underground space, sunlight does not reach it, and lighting equipment, etc. cannot be installed in advance, so the inside of the shield tunnel 101 is almost completely dark. The unmanned exploration system 1 enables the autonomous flight and exploration of the unmanned aerial vehicle 3 with the shield tunnel 101 under the above conditions as a target space.

2(a), the unmanned exploration system 1 includes an unmanned aerial vehicle 3 that flies inside the shield tunnel 101, and an exploration support terminal 5 used by an exploration operator near the entrance of the shield tunnel 101. In this embodiment, a starting point P0 is set near the bottom of the shaft 103 as the exploration base, and the unmanned aerial vehicle 3 takes off and lands at this starting point P0. For example, a pedestal for the unmanned aerial vehicle 3 to take off and land may be installed at the starting point P0. The exploration support terminal 5 is, for example, a computer system used near the starting point P0, and is connected to the general control unit 13 (described later) of the unmanned aerial vehicle 3 by wire or wirelessly to exchange information. The exploration support terminal 5 includes, for example, a personal computer operated by an exploration operator near the starting point P0, and a remote control controller that remotely controls the flight of the unmanned aerial vehicle 3. Alternatively, the function of the remote control controller may be incorporated in the personal computer.

The unmanned aerial vehicle 3 comprises a drone 11 having a power source for flight, and various types of loaded equipment mounted on the drone 11. The loaded equipment includes a general control unit 13, a scanning device 15, an optical camera 17, a gas sensor 19, and other sensors 21. The drone 11 comprises, for example, multiple propellers that generate thrust, a motor that drives each propeller, and a battery that is a power source for the motor. A commercially available known drone may be used as such a drone 11.

The general control unit 13 is composed of a relatively small computer system that can be mounted on the drone 11, for example. The general control unit 13 is connected to a flight control computer (not shown) equipped in the drone 11 to send and receive signals, and can control the flight of the drone 11 via the flight control computer. The general control unit 13 also controls the operation of the scanning device 15, the optical camera 17, the gas sensor 19, and other sensors 21. The general control unit 13 has an information storage unit 13a that stores information necessary for exploring the shield tunnel 101, and a communication unit 13b for communicating with the exploration support terminal 5 via wired or wireless communication.

2(b), the overall control unit 13 is physically configured as a computer system including a CPU 211, RAM 212 and ROM 213 as main storage devices, an auxiliary storage device 214 such as a silicon drive or hard disk, an input device 215 such as a button or switch that accepts user operations, an output device 216 such as a display or indicator lamp that displays information externally, and a communication module 217 that is a data transmission/reception device. Each function of the overall control unit 13 described in this embodiment is realized by loading specific computer software onto hardware such as the CPU 211 and RAM 212, operating the communication module 217, input device 215, and output device 216 under the control of the CPU 211, and reading and writing data from and to the RAM 212 and the auxiliary storage device 214.

The scanning device 15 of the unmanned aerial vehicle 3 is a device known as a 3D LiDAR (3D light detection and ranging) device, which emits invisible laser light into the surroundings while the unmanned aerial vehicle 3 is flying and receives the laser light reflected from the surrounding objects, thereby measuring the distance to each point on the object. The scanning device 15 scans the entire circumference horizontally within a range of ±15° up and down from the horizontal plane, and collects distance information to each point on the object (for example, the inner wall surface of the shield tunnel 101) within this scanning range. Point cloud information, which is a collection of such points, is transmitted to the general control unit 13 and stored and accumulated in the information storage unit 13a.

Based on the point cloud information from the scanning device 15, the overall control unit 13 recognizes the three-dimensional position and three-dimensional shape of the objects around the unmanned aerial vehicle 3 during flight. That is, the overall control unit 13 can recognize the three-dimensional arrangement of the surrounding objects and the relative three-dimensional position of the unmanned aerial vehicle 3 with respect to the surrounding objects. In this way, the process of simultaneously mapping the objects around the unmanned aerial vehicle 3 and identifying the self-position of the unmanned aerial vehicle 3 is called SLAM (Simultaneous Localization and Mapping) processing. In this embodiment, the scanning device 15 and the overall control unit 13 perform LiDAR-SLAM processing using a 3D LiDAR device using laser light. Since LiDAR-SLAM processing is well known, further detailed description will be omitted. The mapping information obtained by LiDAR-SLAM processing is position information of the point cloud on the objects around the unmanned aerial vehicle 3, and is stored in the information storage unit 13a of the overall control unit 13.

Here, as shown in FIG. 3(a), three physical targets 31 for calibrating the self-position of the unmanned aerial vehicle 3 are installed around the starting point P0. The targets 31 are, for example, spherical with a diameter of about 10 to 20 cm, and as shown in FIG. 3(b), each target 31 is attached with a leg 31a for adjusting the height position. All targets 31 are installed within the scanning range of the scanning device 15 of the unmanned aerial vehicle 3 when it lands at the starting point P0. Each target 31 is fixed in position within the shaft 103, for example, by fixing the leg 31a to the bottom surface of the shaft 103. The leg 31a may be omitted and the target 31 may be installed directly on the bottom wall surface or side wall of the shaft 103. In order to facilitate detection of the target 31 by the scanning device 15, the spherical surface of the target 31 is made of a retroreflective material that retroreflects the laser light of the scanning device 15.

The three installed targets 31 have been surveyed in advance, and the three-dimensional absolute coordinates of each target 31 are known. The absolute coordinates of each target 31 obtained by the above surveying are stored in advance in the information storage unit 13a of the unmanned aerial vehicle 3's general control unit 13. The scanning device 15 of the unmanned aerial vehicle 3 scans each target 31, and the general control unit 13 can obtain the relative positions of the three targets 31 with respect to the unmanned aerial vehicle 3. The general control unit 13 can then recognize the absolute position of the unmanned aerial vehicle 3 based on a comparison between the relative positions of each target 31 obtained by the above scan and the absolute coordinates of each target 31 stored in advance in the information storage unit 13a.

The "absolute position" of the unmanned aerial vehicle 3 includes the absolute three-dimensional coordinates of the unmanned aerial vehicle 3 and the absolute three-dimensional orientation (attitude) of the unmanned aerial vehicle 3. In order to improve the recognition accuracy of the absolute position of the unmanned aerial vehicle 3, the three targets 31 are preferably placed at the vertices of a triangle surrounding the starting point P0 in a planar view, and it is preferable to make the triangle large. In addition, in order to facilitate detection by the scanning device 15, it is preferable that the height position of the target 31 is approximately the same height position as the scanning device 15 when the unmanned aerial vehicle 3 lands at the starting point P0. In addition, if the size of the target 31 is too small, it becomes difficult for the scanning device 15 to detect it, and if the size of the target 31 is too large, the positional accuracy detected by the scanning device 15 decreases. From this perspective, it is preferable that the size of the target 31 is 10 to 20 cm in diameter, as described above. In addition, the number of targets 31 to be installed may be three or more, but four or more targets 31 may be installed to improve the recognition accuracy of the absolute position.

The unmanned aerial vehicle 3 is further equipped with sensors that acquire environmental information within the shield tunnel 101 during flight. As shown in FIG. 2(a), as examples of such sensors, the unmanned aerial vehicle 3 of this embodiment is equipped with the aforementioned optical camera 17, gas sensor 19, and other sensors 21. The environmental information acquired by these sensors is sent to the overall control unit 13, and the overall control unit 13 transmits the environmental information to the exploration support terminal 5 via the communication unit 13b.

The optical camera 17 is, for example, a 360° camera, and captures video of the surroundings of the unmanned aerial vehicle 3 using visible light. As mentioned above, since the inside of the shield tunnel 101 is a dark place, the unmanned aerial vehicle 3 is equipped with a lighting device (for example, a high-brightness LED video light) that projects light toward the surroundings to enable video capture by the optical camera 17. This optical camera 17 captures video of the surroundings of the unmanned aerial vehicle 3, and can obtain information on the appearance of the inner surface of the shield tunnel 101, for example. A commercially available 360° camera or the like may be used as the optical camera 17.

The gas sensor 19 is a sensor that measures the concentration of a specific gas component (e.g., a toxic gas) contained in the air near the unmanned aerial vehicle 3. Examples of gas components whose concentrations are measured by the gas sensor 19 include oxygen, carbon monoxide, flammable gases, and hydrogen sulfide. The other sensors 21 are sensors for acquiring other environmental information near the unmanned aerial vehicle 3. Examples of the other sensors 21 include a temperature sensor, a humidity sensor, a dust concentration sensor, an illuminance sensor, a wind speed sensor, a noise sensor, and a vibration sensor.

The physical configuration of the unmanned aerial vehicle 3 is as follows. As shown in FIG. 4, the drone 11 is equipped with four propellers 11a arranged at the upper end. Furthermore, the unmanned aerial vehicle 3 is equipped with a pair of landing legs 23 extending downward from the drone 11 at symmetrical positions on the left and right of the aircraft. The landing legs 23 may be provided as part of a commercially available drone 11. A beam section 27 extending in the left and right direction of the aircraft is provided so as to span between these landing legs 23. A rectangular parallelepiped-shaped general control unit 13 is fixed to the central upper surface of this beam section 27, and a cylindrical scanning device 15 of the same size is attached to the upper surface of the general control unit 13. That is, the general control unit 13 and the scanning device 15 are arranged vertically stacked in the central part below the drone 11. In this way, the general control unit 13 and the scanning device 15, which are the main heavy objects, are arranged in the central part below the drone 11, thereby enabling the unmanned aerial vehicle 3 to fly stably. In addition, the optical camera 17, lighting device, gas sensor 19, and other sensors 21, which are relatively small and lightweight, are not shown in the figure, but may be appropriately attached to the upper or lower surface of the drone 11 or the landing legs 23, etc., taking into consideration the weight balance.

Next, the exploration method in the shield tunnel 101 using the unmanned exploration system 1 as described above will be described with reference to Figures 1, 5, and 6. Figures 5 and 6 are flowcharts of the exploration work in the shield tunnel 101 using the exploration method. The operation of the unmanned aerial vehicle 3 and the like in the following description is realized by the general control unit 13 and the exploration support terminal 5 executing a predetermined flight control program. In this exploration work, the movement method of the unmanned aerial vehicle 3 in the shield tunnel 101 is such that the first step, the second step, ..., the mth step (m = 2, 3, ...) as described below are executed in order to gradually add the movable areas A1, A2, ..., Am toward the moving direction (i.e., toward the face side of the shield tunnel 101). Hereinafter, the face side of the shield tunnel 101 may be referred to as "forward in the moving direction" or simply "forward", and the entrance side of the shield tunnel 101 may be referred to as "rearward in the moving direction" or simply "rear". The value of m above is appropriately determined taking into consideration conditions such as the range of the unmanned aerial vehicle 3.

More details will be given later.
In the first step, mapping information of an area A1 ahead of a starting point P0 is acquired, and a waypoint P1 is set within the area A1;
In the second step, mapping information for an area A2 ahead of the area A1 is acquired, and a waypoint P2 is set within the area A2.
In the third step, mapping information for an area A3 ahead of the area A2 is acquired, and a waypoint P3 is set within the area A3.
…
In the mth step, mapping information for an area Am further ahead of the area A(m-1) is acquired, and a waypoint Pm is set within the area Am.

All of areas A1 to Am are part of the shield tunnel 101, and waypoints P1 to Pm are virtual points set within areas A1 to Am as flight waypoints for the unmanned aerial vehicle 3 and recognized by the general control unit 13. Waypoints P1 to Pm are manually selected by the exploration operator and set at positions close to the forefront of areas A1 to Am. Note that waypoints P1 to Pm are flight waypoints for the unmanned aerial vehicle 3 and are not landing points for the unmanned aerial vehicle 3, so there is no need to select points suitable for landing.

The first, second, ..., mth steps are explained in detail below. Here, one of the first, second, ..., mth steps is generalized as the kth step (k = 1 to m), and the explanation is divided into the case where k = 1 (first step) and the case where k = 2 to m.

[First step (when k = 1)]
As shown in FIG. 5, in the first step, the unmanned aerial vehicle 3 is first placed at the starting point P0 (step S302 in FIG. 5). At this time, the general control unit 13 of the unmanned aerial vehicle 3 detects three surrounding targets 31 by LiDAR-SLAM processing using the scanning device 15, and recognizes the exact absolute position of the unmanned aerial vehicle 3 (step S304: calibration). After that, the general control unit 13 starts the drone 11, and the unmanned aerial vehicle 3 is raised, for example, vertically at the starting point P0 and then lowered again to land again at the starting point P0 (step S306). Here, the unmanned aerial vehicle 3 rises, for example, by about 5 m so as to reach a position higher than the ceiling of the entrance of the shield tunnel 101. The above calibration (step S304) may be performed during this ascent/descent (step S306). During the above ascent/descent (step S306), the scanning device 15 continuously scans surrounding objects, and the overall control unit 13 performs mapping of the surroundings of the starting point P0 by LiDAR-SLAM processing. Then, when the unmanned aerial vehicle 3 lands again at the starting point P0, the mapping information of the surroundings of the starting point P0 is stored in the information storage unit 13a (step S308).

In the LiDAR-SLAM process described above, the laser light from the scanning device 15 reaches the back of the shield tunnel 101 through the entrance of the shield tunnel 101, so mapping information can be obtained up to the area ahead of the entrance of the shield tunnel 101. It is said that the distance at which objects can be detected by a typical 3D LiDAR device is about 100 m. Therefore, it is considered that mapping information can be obtained up to a position, for example, about 80 m ahead of the entrance of the shield tunnel 101 (about 100 m ahead of the starting point P0). In this way, in the first step, mapping information for "area A1," which is the area up to a position, for example, about 80 m ahead of the entrance of the shield tunnel 101, is obtained and stored in the information storage unit 13a.

This mapping information is acquired by the scanning device 15 as information indicating the relative position with respect to the unmanned aerial vehicle 3. However, since the overall control unit 13 recognizes the absolute position of the unmanned aerial vehicle 3 through calibration (step S304) using the aforementioned target 31, the overall control unit 13 converts the acquired mapping information into mapping information indicated by absolute position and stores it in the information storage unit 13a. Similarly, all of the mapping information acquired by the scanning device 15 thereafter, the position information of waypoints P1 to Pm, flight route information, current position information of the unmanned aerial vehicle 3, and other position information are all recognized by the overall control unit 13 as information indicated by absolute position.

After the unmanned aerial vehicle 3 lands again at the starting point P0 and stops, the exploration operator connects the exploration support terminal 5 to the general control unit 13 of the unmanned aerial vehicle 3 by wire or wirelessly, and reads out the mapping information of area A1 stored in the information storage unit 13a to the exploration support terminal 5 (step S310). In addition, after the mapping information of area A1 read out to the exploration support terminal 5 is stored in the exploration support terminal 5, it is deleted from the information storage unit 13a, thereby saving storage capacity of the information storage unit 13a.

A specific flight plan program is executed on the exploration support terminal 5, and the mapping information of area A1 that has been read out is displayed as a visual image showing a cloud of points on an object (e.g., the inner wall surface of the shield tunnel 101) that exists within area A1. While referring to this image, the exploration operator selects an arbitrary point in the forward vicinity within area A1 and performs an operation to provisionally set it as waypoint P1 (step S312).

Next, the exploration operator executes a flight simulator function included in the flight plan program, and executes a flight simulation on the exploration support terminal 5 based on the mapping information of the area A1 and the position information of the waypoint P1 stored in the exploration support terminal 5 (step S314). In this flight simulation, it is confirmed that the unmanned aerial vehicle 3 can safely fly round trip from the starting point P0 to the waypoint P1 and return based on the mapping information of the area A1. If the simulation shows that the unmanned aerial vehicle 3 cannot safely fly round trip, the simulation is repeated while making adjustments such as changing the waypoint P1 and resetting the flight route as necessary. Finally, the waypoint P1 that allows safe round trip flight from the starting point P0 and the flight route F1 from the starting point P0 to the waypoint P1 are determined. The determined position information of the waypoint P1 and the information of the flight route F1 are stored in the exploration support terminal 5, and are also transmitted to the general control unit 13 of the unmanned aerial vehicle 3 and stored in the information storage unit 13a. This completes the first step.

[kth step (when k = 2 to m)]
Following the above-mentioned first step, the second step, the third step, ..., the mth step are executed in order. For these second to mth steps, substantially the same steps are repeated (see steps S401, S418, and S420 in FIG. 6). In the following explanation, detailed explanations of processes that are the same or equivalent to the above-mentioned first step will be omitted as appropriate.

At the start of the kth step,
In the first step, a waypoint P1 and a flight route F1 within the area A1 have been set,
In the second step, a waypoint P2 and a flight route F2 within the area A2 have been set,
In the third step, a waypoint P3 and a flight route F3 in the area A3 have been set,
…
In the (k-1)th step, waypoint P(k-1) and flight route F(k-1) within area A(k-1) have been set.
The position information of each of the waypoints P1 to P(k-1) and the flight route information F1 to F(k-1) are stored in the exploration support terminal 5 and the information storage unit 13a of the overall control unit 13. The exploration support terminal 5 further stores mapping information of the areas A1 to A(k-1). As described above, the position information of the waypoints P1 to P(k-1), the waypoints P1 to P(k-1), and the mapping information of the areas A1 to A(k-1) are stored as information indicating absolute positions.

In the kth step, the unmanned aerial vehicle 3 is placed at the starting point P0, and flies round trip from the starting point P0 to the waypoint P(k-1) along the flight route F1 to F(k-1) under the control of the overall control unit 13. This round trip flight is performed autonomously by the unmanned aerial vehicle 3, and the exploration operator in the vicinity of the starting point P0 does not need to operate the unmanned aerial vehicle 3 during flight, except in an emergency. The procedure for this round trip flight is as follows.

As shown in FIG. 6, first, the unmanned aerial vehicle 3 is placed at the starting point P0 (step S402 in FIG. 6), and then the general control unit 13 starts the drone 11 to take off the unmanned aerial vehicle 3. Before or immediately after the takeoff, the general control unit 13 of the unmanned aerial vehicle 3 detects three targets 31 (FIG. 3) by LiDAR-SLAM processing and recognizes the absolute position of the unmanned aerial vehicle 3 (step S404: calibration). After that, the unmanned aerial vehicle 3 flies under the control of the general control unit 13, and flies autonomously to waypoint P(k-1) based on the position information of waypoints P1 to P(k-1) and flight route information F1 to F(k-1) stored in the information storage unit 13a (step S406). Here, when k=3 to m, the unmanned aerial vehicle 3 reaches waypoint P(k-1) by passing through waypoints P1 to P(k-1) in order.

During this flight, the scanning device 15 continuously scans for surrounding objects, and the overall control unit 13 maps the area around the unmanned aerial vehicle 3 using LiDAR-SLAM processing. The current position information of the unmanned aerial vehicle 3, which is necessary for flight control by the overall control unit 13, is recognized in real time by a self-position estimation process that combines LiDAR-SLAM processing and estimation processing using the IMU (Inertial Measurement Unit) equipped on the drone 11.

During flight, the optical camera 17 of the unmanned aerial vehicle 3 continuously captures video of the surroundings, the gas sensor 19 measures the concentration of gas components, and other sensors 21 continuously collect other environmental information. The general control unit 13 then associates the collected environmental information with the current position information of the unmanned aerial vehicle 3 and continuously transmits it to the exploration support terminal 5 in real time via the communication unit 13b (step S407). The exploration operator can know the environment inside the shield tunnel 101 by analyzing each of these pieces of environmental information received by the exploration support terminal 5. For example, the soundness of the segments of the shield tunnel 101 can be known based on the video of the inner wall surface of the shield tunnel 101 captured by the optical camera 17. Also, for example, the presence or absence of toxic gases inside the shield tunnel 101 can be known from the concentration information of gas components measured by the gas sensor 19.

Note that the above-mentioned capturing of video of the surroundings by the optical camera 17, measuring the concentration of gas components by the gas sensor 19, and collecting other environmental information by the other sensors 21 (step S407) may be performed only in the final step, the mth step. In this way, in the first to (m-1)th steps, there is no need to load the optical camera 17, gas sensor 19, and other sensors 21 onto the unmanned aerial vehicle 3, making it possible to reduce the weight of the unmanned aerial vehicle 3 and extend its range.

During the flight of the unmanned aerial vehicle 3, as described above, the scanning device 15 continuously scans surrounding objects, and the general control unit 13 performs mapping of the surroundings by LiDAR-SLAM processing. Then, when the unmanned aerial vehicle 3 reaches waypoint P(k-1), mapping information of the surroundings of the waypoint P(k-1) is stored in the information storage unit 13a. That is, at this point, mapping information of "area Ak", which is the area ahead of waypoint P(k-1), is stored in the information storage unit 13a (step S408). For example, the area Ak is about 100 m further ahead of waypoint P(k-1). However, for example, in the case where the laser light of the scanning device 15 does not reach sufficiently ahead of waypoint P(k-1) at a bend in the shield tunnel 101, the area Ak will be narrower than the above.

Then, under the control of the overall control unit 13, the unmanned aerial vehicle 3 autonomously flies to the starting point P0 and returns along a flight route that follows the outward route in reverse, based on the position information of waypoints P1 to P(k-1) and flight route information F1 to F(k-1) stored in the information storage unit 13a (step S409).

In addition, while the unmanned aerial vehicle 3 is flying, a processing program is also executed in parallel to execute the following measures for safe flight. For example, if an obstacle is detected ahead in the direction of movement of the unmanned aerial vehicle 3 by LiDAR-SLAM processing, the unmanned aerial vehicle 3 will automatically avoid the obstacle and fly away. This processing makes it possible to deal with the case where an obstacle that did not exist up until the (k-1)th step appears later. Alternatively, if the obstacle is unavoidable and is not removed for a predetermined time, the unmanned aerial vehicle 3 will automatically return to the starting point P0. In addition, if the battery of the drone 11 is expected to run out (for example, if the remaining battery capacity falls below a predetermined value), the unmanned aerial vehicle 3 will automatically return to the starting point P0 or land on the spot.

In addition, if dust, fog, smoke, or the like occurs around the unmanned aerial vehicle 3, this may hover the unmanned aerial vehicle 3, as this may interfere with LiDAR-SLAM processing. Alternatively, if the dust, etc. does not subside within a specified time, the unmanned aerial vehicle 3 automatically returns to the starting point P0. In addition, in the event of an emergency, the exploration operator at the starting point P0 can operate the remote control of the exploration support terminal 5 to manually remotely control the unmanned aerial vehicle 3.

After the unmanned aerial vehicle 3 that has returned from waypoint P(k-1) lands and stops at the starting point P0, the exploration operator connects the exploration support terminal 5 to the unmanned aerial vehicle 3's general control unit 13 via the communication unit 13b by wire or wirelessly, and reads out the mapping information of area Ak stored in the information storage unit 13a to the exploration support terminal 5 (step S410). Then, as in the first step, the exploration operator executes a flight plan program on the exploration support terminal 5 to provisionally set waypoint Pk within area Ak (step S412).

The exploration operator then performs a flight simulation on the exploration support terminal 5 in the same manner as in the first step (step S414). This flight simulation adjusts the waypoint Pk as necessary, and adjusts the flight route Fk from waypoint P(k-1) to waypoint Pk, and determines the waypoint Pk and flight route Fk that allows for a safe round-trip flight from the starting point P0. The determined position information of the waypoint Pk and information on the flight route Fk are stored in the exploration support terminal 5, and are also transmitted to the general control unit 13 of the unmanned aerial vehicle 3 and stored in the information storage unit 13a (step S416). This completes the kth step.

After the kth step is completed, if the kth step is the final step, the mth step (step S418), the exploration work inside the shield tunnel 101 is terminated, and if the mth step has not yet been reached, the value of k is counted up (step S420) and the process returns to step S402. Note that in the final step, the mth step, the processes of steps S410 to S416 for setting the next waypoint Pm and the flight route Fm may be omitted. Also, in the final step, the mth step, the process in which the information storage unit 13a stores the mapping information of the area Am when the unmanned aerial vehicle 3 reaches the waypoint P(m-1) (part of step S408) may be omitted.

Next, we will explain the effects of the method for exploring the shield tunnel 101 using the unmanned exploration system 1 as described above.

According to the above exploration method, the areas A1, A2, ..., Am in which the unmanned aerial vehicle 3 can move are gradually added forward in the direction of movement, for example, by about 30 to 80 m at a time, and the exploration area by the unmanned aerial vehicle 3 can be gradually expanded toward the face side of the shield tunnel 101. At this time, the unmanned aerial vehicle 3 in the kth step flies to waypoint P(k-1) based on the flight route F1 to F(k-1) set up to the (k-1)th step. In particular, when k = 3 to m, the unmanned aerial vehicle 3 flies so as to pass through waypoints P1 to P(k-1) set up to the (k-1)th step in order to reach waypoint P(k-1). Here, the above waypoints P1 to P(k-1) and flight routes F1 to F(k-1) have been confirmed to enable safe round-trip flight of the unmanned aerial vehicle 3 through flight simulation up to the (k-1) step, so the unmanned aerial vehicle 3 can safely fly round-trip from the starting point P0 to the waypoint P(k-1).

In addition, the current position information of the unmanned aerial vehicle 3, which is necessary for flight control by the overall control unit 13, is recognized in real time by a self-position estimation process that combines LiDAR-SLAM processing and estimation processing by the IMU equipped in the drone 11. Therefore, GNSS signals are not required for flight control of the unmanned aerial vehicle 3, and it is also not necessary to install signs that the unmanned aerial vehicle 3 can read or equipment that communicates with the unmanned aerial vehicle 3 in advance inside the shield tunnel 101, allowing the unmanned aerial vehicle 3 to fly autonomously.

Therefore, the above exploration method can be used to ideally explore spaces where GNSS signals cannot be used and where people cannot enter in advance, such as the shielded tunnel 101. In addition, since the scanning device 15 of the unmanned aerial vehicle 3 is a LiDAR device that uses laser light, the above exploration method can be used to ideally explore dark places, such as the shielded tunnel 101.

In addition, at least three targets 31 with known absolute coordinates are installed near the starting point P0, and the unmanned aerial vehicle 3's general control unit 13 can calibrate the absolute position of the unmanned aerial vehicle 3 at the beginning of each process (first to mth processes). Therefore, the unmanned exploration system 1 can consistently recognize the position information of waypoints P1 to Pm in each process (first to mth processes), mapping information of areas A1 to Am, information on flight routes F1 to Fm, and current position information of the unmanned aerial vehicle 3 as absolute position information.

If the above-mentioned various position information is recognized as relative position information with respect to the unmanned aerial vehicle 3 rather than absolute position information, it is necessary to accurately install the unmanned aerial vehicle 3 at the start of each process (first to mth processes) in the same position and in the same attitude at the start point P0. That is, in this case, the positions of the waypoints P1 to Pm are recognized as relative positions based on the installation position and installation attitude of the unmanned aerial vehicle 3 at the start of each process, so if there is an error in the installation position and installation attitude of the unmanned aerial vehicle 3 at the start point P0, the subsequent flight route will be shifted according to the error. In particular, the error in the installation attitude of the unmanned aerial vehicle 3 can cause a larger deviation the longer the flight route. In contrast, according to the exploration method of this embodiment, the positions of the waypoints P1 to Pm are recognized as absolute position information regardless of the position and attitude of the unmanned aerial vehicle 3, so it is not necessary to install the unmanned aerial vehicle 3 with high positional accuracy at the start of each process (first to mth processes), and as a result, the workload is reduced.

The present invention can be implemented in various forms, including the above-described embodiment, with various modifications and improvements based on the knowledge of those skilled in the art. It is also possible to configure modified examples by utilizing the technical matters described in the above-described embodiment. The configurations of each embodiment, etc. may be used in appropriate combination.

For example, it is not essential that the unmanned aerial vehicle 3 fly by passing through waypoints P1 to P(k-1) in order in the kth step (where k = 3 to n). For example, the unmanned aerial vehicle 3 may fly along a flight route from the starting point P0 to waypoint P(k-1) without necessarily passing through waypoints P1 to P(k-2). In this case, the flight route as described above may be set in the (k-1)th step and stored in the general control unit 13.

In addition, in the embodiment, after the unmanned aerial vehicle 3 returns to the starting point P0, mapping information is read from the overall control unit 13 to the exploration support terminal 5, and waypoint setting, flight route setting, and flight simulation are performed on the exploration support terminal 5. However, this procedure is not essential, and for example, the exploration operator may connect necessary equipment such as an operation input device (e.g., a keyboard) and a display device (e.g., a display monitor) to the overall control unit 13, and a flight plan program for setting waypoints, setting flight routes, and performing flight simulation may be executed directly on the overall control unit 13.

In addition, the general control unit 13 of the embodiment performs LiDAR-SLAM processing using a 3D LiDAR device (scanning device 15) as the SLAM processing of the unmanned aerial vehicle 3, but instead of this, for example, Visual-SLAM processing using images from the optical camera 17 may be performed. Furthermore, the scanning device 15 performs scanning within a scanning range of ±15° in the vertical direction from the horizontal plane, but this entire scanning range may be tilted with respect to the horizontal plane. For example, if the inclination of the scanning range is adjusted to correspond to the inclination of the tunnel when flying through an inclined tunnel, the detection sensitivity of the inner wall surface of the tunnel can be increased more easily than with a method of flying the drone 11 in an inclined attitude.

In addition, although the drone 11 in this embodiment is motor-driven, a drone with a gasoline engine or a drone with a hybrid engine may be used instead. However, a motor-driven drone 11 is preferable in that it reduces vibration compared to a gasoline engine or a hybrid engine. In addition, the unmanned aerial vehicle 3 may be capable of landing on the surface of water by providing a float at the lower end of the landing leg 23 of the unmanned aerial vehicle 3, for example.

In addition, in the exploration method and movement method of the present invention, an unmanned mobile body moving on the ground may be used instead of the unmanned aerial vehicle 3 that flies in the target space. For example, the unmanned mobile body may be a track-type running device or a quadruped walking robot equipped with the various loading devices (general control unit 13, scanning device 15, etc.) described above instead of a drone 11. However, the unmanned aerial vehicle 3 is preferable to a track-type running device or a quadruped walking robot in that it can move regardless of the condition of the bottom surface of the target space.

The exploration method and movement method of the present invention can be applied to various applications other than collecting environmental information about the shield tunnel 101. For example, the exploration method and movement method of the present invention can be applied to automatic inspection of mountain tunnels and underground tunnels, exploration and investigation of dangerous areas (areas where toxic gases or areas at risk of collapse exist), inspection of joint structures, etc., inspection of ship ballast tanks where there is a possibility of the presence of hydrogen sulfide or oxygen deficiency, patrol inspection of construction site buildings and underground structures, forested areas in mountainous areas where it is difficult to receive GNSS signals, accident and disaster investigations, measurement of the three-dimensional shape of tunnel faces, etc.

3...Unmanned aerial vehicle (unmanned mobile vehicle), 15...Scanning device, 17...Optical camera (sensors), 19...Gas sensor (sensors), 21...Other sensors, 23...Landing legs, 27...Beam, 31...Target, Ak, A1 to Am...Area, P0...Start point, Pk, P1 to Pm...Waypoint, Fk, F1 to Fm...Flight route, 101...Shield tunnel (target space).

Claims (7)

A method for moving an unmanned moving body in a predetermined target space extending forward in a moving direction from a predetermined starting point, comprising the steps of:
The method includes a first step, a second step, a third step, ..., in which a region in the target space is mapped part by part,
In the kth step (where k=2, 3, ...),
A method for moving an unmanned mobile body, in which the unmanned mobile body is installed at the starting point, and then the unmanned mobile body moves to the (k-1)th waypoint stored in the (k-1)th step by SLAM processing, maps a kth area in front of the (k-1)th waypoint, and returns to the starting point, and the kth waypoint set within the mapped kth area is stored in the unmanned mobile body. At least three physical targets detectable by the SLAM processing of the unmanned vehicle are fixed near the starting point;
In the kth step (where k=1, 2, . . . ),
The method for moving an unmanned mobile body according to claim 1 , wherein the unmanned mobile body recognizes its own position based on the position of the physical target detected by the SLAM process in the vicinity of the starting point. In the kth step (where k=1, 2, . . . ),
A method for moving an unmanned mobile body as described in claim 1 or 2, in which a movement simulation of the unmanned mobile body is performed based on mapping information of the kth region stored by the unmanned mobile body that has returned to the starting point, and a point at which the movement simulation determines that the unmanned mobile body can move back and forth from the starting point is set as the kth waypoint. In the kth step (where k=3, 4, . . . ),
The method for moving an unmanned moving body described in any one of claims 1 to 3, wherein the unmanned moving body moves from the starting point to the (k-1)th waypoint along a movement route that passes through the first to (k-1)th waypoints in order from the starting point. The method for moving an unmanned moving body according to any one of claims 1 to 4, wherein the unmanned moving body is an unmanned aerial vehicle capable of flying within the target space and equipped with a scanning device for the SLAM processing. The method for moving an unmanned mobile body according to claim 5, wherein the unmanned aerial vehicle comprises a main body having a power source, a pair of landing legs extending downward from the main body, and a beam section that spans between the landing legs below the main body, and the scanning device is installed on the beam section. The unmanned moving body, which is equipped with sensors for acquiring environmental information within the target space, moves within the target space by the method for moving an unmanned moving body according to any one of claims 1 to 6,
An exploration method, in which the environmental information obtained by the sensors is collected while the unmanned mobile body is moving within the target space.

Images