JP7799554B2 - Robot control system, transport system, and robot control method - Google Patents

Description

The present invention relates to technology for correcting a robot's self-position.

A known method for estimating the self-location of an automatic guided vehicle (AGV) is to combine multiple self-location estimation methods to improve accuracy and robustness. For example, one method combines the Visual SLAM (SLAM: Simultaneous Localization and Mapping) method (first self-location estimation method) that uses image feature points with the Wheel Odometry (hereinafter referred to as "odometry") method (second self-location estimation method) that uses wheel rotation information (encoder information).

Patent Document 1 discloses a self-location estimation method in which a calculation accuracy map is created in advance for a first self-location obtained by image processing-based self-location estimation processing and a second self-location obtained by odometry-based self-location estimation processing, and the calculation accuracy recorded in the calculation accuracy map is used to determine the combination ratio of the first self-location and the second self-location.

On the other hand, Figure 5 of Patent Document 1 discloses a method for estimating self-location using only the second self-location when a predetermined number of landmarks cannot be detected, i.e., when the first self-location cannot be obtained due to an error.

Patent Document 2 and Non-Patent Document 1 disclose the basic functionality of Visual SLAM using image feature points, specifically the functionality of detecting a group of image feature points from successive camera images and tracking this group of feature points across multiple frames to estimate self-location while creating a map. Furthermore, they also disclose a functionality of estimating self-location with high accuracy by matching camera images against a map created using the basic functionality of Visual SLAM.

JP 2021-96731 A JP 2017-146952 A

Shinya Sumikura, Mikiya Shibuya, and Ken Sakurada, "OpenVSLAM: A Versatile Visual SLAM Framework", Open Source Software Competition, October 21-25, 2019, Nice

Patent Document 1 discloses a method for estimating a self-location using only a second self-location when the first self-location cannot be obtained due to an error. The second self-location obtained by odometry-based self-location estimation processing is susceptible to factors such as friction and unevenness on the road surface (or floor surface) and wheel wear, and is less accurate in position estimation than the first self-location obtained by image processing-based self-location estimation processing.

As a result, when the second self-location is used, there is an issue in that errors accumulate compared to the first self-location that would have been obtained if no errors had occurred.

The present invention was made in consideration of the above-mentioned problems, and aims to improve the accuracy of estimating self-position using information calculated from the second sensor when controlling a robot having different types of first and second sensors and the first self-position calculated from the first sensor is unavailable.

The present invention is a robot control system in which a server having a processor and memory estimates the position of a mobile robot, the mobile robot having a first sensor unit that acquires first sensing information from the traveling state of the mobile robot, and a second sensor unit that has a sensor different from the first sensor unit and acquires second sensing information from the traveling state of the mobile robot, the server acquiring first traveling state information including a first position and a first attitude at the first position obtained from the first sensing information, and a second velocity and second speed of the mobile robot obtained from the second sensing information. The system has a storage unit that stores correspondence information indicating the correspondence between the first position and second running state information including angular velocity information, in association with the first position, and a calculation unit that is capable of calculating an estimated position of the mobile robot based on the first running state information and the second running state information, wherein the calculation unit calculates the first position and the first orientation information based on the first sensing information acquired from the first sensor unit, and if an error occurs when calculating the first position, calculates the estimated position of the mobile robot based on the second running state information acquired from the second sensing information and the correspondence information.

Therefore, even if the first position cannot be obtained from the first sensing information of the first sensor unit due to an error, the present invention can calculate the position of the mobile robot with the same accuracy as the first position using the second sensing information of the second sensor unit and the velocity ratio and angular velocity ratio of the correspondence information.

Details of at least one implementation of the subject matter disclosed herein are set forth in the accompanying drawings and the description below. Other features, aspects, and advantages of the disclosed subject matter will become apparent from the following disclosure, drawings, and claims.

FIG. 2 illustrates a motion model of a robot according to the first embodiment of the present invention. FIG. 2 illustrates the relationship between the trajectory of the robot and its own position according to the first embodiment of the present invention. FIG. 10 illustrates the velocity and angular velocity of the robot according to the first embodiment of the present invention. FIG. 1 is a block diagram illustrating an example of hardware of a robot control system according to a first embodiment of the present invention. FIG. 1 is a perspective view of a robot according to a first embodiment of the present invention. FIG. 1 is a side view of a robot according to a first embodiment of the present invention. FIG. 2 is a block diagram illustrating an example of software for the robot control system according to the first embodiment of the present invention. 10 is a flowchart illustrating an example of a self-position correction process according to the first embodiment of the present invention. 10 is a flowchart illustrating an example of a database search process according to the first embodiment of the present invention. FIG. 10 illustrates the first embodiment of the present invention and is a diagram illustrating an example of correction information in the case of traveling straight. FIG. 10 illustrates the first embodiment of the present invention and is a diagram illustrating an example of correction information in the case of a right turn. 10 is a flowchart illustrating an example of correction information in a database according to the first embodiment of the present invention. FIG. 10 illustrates the first embodiment of the present invention and is a diagram illustrating an example of correction information for each movement type in the case of turning right.

Embodiments of the present invention will be described below with reference to the drawings. The examples are illustrative only and have been omitted or simplified as appropriate for clarity of explanation. The present invention can also be implemented in various other forms. Unless otherwise specified, each component may be singular or plural. The position, size, shape, range, etc. of each component shown in the drawings may not represent the actual position, size, shape, range, etc., in order to facilitate understanding of the invention. Therefore, the present invention is not necessarily limited to the position, size, shape, range, etc. disclosed in the drawings.

As examples of various types of information, expressions such as "table," "list," and "queue" may be used, but various types of information may also be expressed using other data structures. For example, various types of information such as "XX table," "XX list," and "XX queue" may also be expressed as "XX information." When describing identification information, expressions such as "identification information," "identifier," "name," "ID," and "number" are used, but these are interchangeable.

When there are multiple components with the same or similar functions, they may be described using the same reference numeral with different subscripts. Also, when there is no need to distinguish between these multiple components, the subscripts may be omitted.

In the embodiments, processing performed by executing a program may be described. Here, a computer executes the program using a processor (e.g., a CPU or GPU) and performs the processing defined in the program using storage resources (e.g., memory) and interface devices (e.g., communication ports). Therefore, the entity performing the processing by executing the program may be the processor. Similarly, the entity performing the processing by executing the program may be a controller, device, system, computer, or node having a processor. The entity performing the processing by executing the program may be any computing unit, and may include a dedicated circuit that performs specific processing. Here, a dedicated circuit is, for example, an FPGA (Field Programmable Gate Array), ASIC (Application Specific Integrated Circuit), or CPLD (Complex Programmable Logic Device).

A program may be installed on a computer from a program source. The program source may be, for example, a program distribution server or a computer-readable storage medium. If the program source is a program distribution server, the program distribution server may include a processor and storage resources for storing the program to be distributed, and the processor of the program distribution server may distribute the program to be distributed to other computers. Furthermore, in the embodiments, two or more programs may be realized as one program, or one program may be realized as two or more programs.

The present invention allows an autonomously traveling automated guided vehicle to restore its own position with the same accuracy as the first self-position from the driving state calculated from the second sensor, even if the first self-position (self-position 1) calculated from the first sensor cannot be obtained due to an error.

In this embodiment, an automatic guided vehicle (AGV) is referred to as a robot. The first and second sensors are composed of different types of sensors, and the position accuracy calculated from the first sensor is higher than the position accuracy calculated from the second sensor.

The method for restoring the robot's own position will be explained using Figure 1A. Equation 100 shown in Figure 1A is a motion model for a two-wheel differential drive robot 300. The robot 300 moves in a two-dimensional space of X-Y and changes its direction of travel by turning. The position of the robot 300 is indicated by x and y coordinates, its direction by θ, its speed by ν, and its angular velocity by ω.

The above equation 100 takes as input the position and orientation (x(t-1), y(t-1), θ(t-1)), velocity (ν(t)), and angular velocity (ω(t)) of robot 300 at time t = 1, 2, 3, 4, ..., and calculates the position and orientation (x(t), y(t), θ(t)) of robot 300 at the next step (time), i.e., it is an equation that expresses the trajectory of robot 300.

In this embodiment, the variable x and subscript t in equation 100 are written as x(t). Variables y, θ, etc. and subscripts are similarly written as y(t) and θ(t). The subscript (t) may also be omitted.

Figure 1B is a diagram showing the relationship between the trajectory and self-position of robot 300. In the figure, trajectory 101 is the trajectory of self-position 1 calculated from velocity 1 (ν1) and angular velocity 1 (ω1) (110) obtained from the first sensor (described below). Trajectory 102 is the trajectory of self-position 2 calculated from velocity 2 (ν2) and angular velocity 2 (ω2) (111) obtained from the second sensor.

The sensor for self-position 1 is highly accurate, but is prone to errors known as self-position loss, in which the sensor loses track of self-position 1. On the other hand, the sensor for self-position 2 does not generate errors, but is prone to errors in the self-position.

As such, the speed and angular velocity obtained from two types of sensors with different performance and accuracy do not necessarily match, resulting in a discrepancy 103 between the two trajectories. This discrepancy 103 is the cumulative distance resulting from the difference between speed 1 (ν1) and speed 2 (ν2) and the difference between angular velocity 1 (ω1) and angular velocity 2 (ω2), which are input at each step (t = 1, 2, 3, 4, ...).

If the above trajectories 101 and 102 were to be perfectly matched, then, as shown in equation 112 in Figure 1B, the velocity ratio (ν1/ν2) and angular velocity ratio (ω1/ω2) would be multiplied by velocity 2 (ν2) and angular velocity ratio (ω2), respectively, so that velocity 2 (ν2) = velocity 1 (ν1) and angular velocity 2 (ω2) = angular velocity 1 (ω1).

However, in practice, it is difficult to perfectly match trajectories 101 and 102, so we attempt to approximate the trajectories using the following method.

First, the velocity ratio (ν1/ν2) and angular velocity ratio (ω1/ω2) when self-position 1 is obtained normally are linked to the position and orientation of robot 300, velocity 2, and angular velocity 2 (x, y, θ, ν2, ω2), and registered as correction information in the database (hereinafter referred to as DB).

If self-position 1 cannot be obtained due to an error or other reason, the speed ratio and angular velocity ratio closest to the current robot state (x, y, θ, v2, ω2) are searched for and used for correction from the speed ratios and angular velocity ratios registered in the DB. Specifically, the speed 1 and angular velocity 1 that would have been obtained if no error had occurred are calculated by approximating them using the following formulas.

Speed 1 (approximation) = Speed 2 × Speed ratio obtained by searching from the DB Angular velocity 1 (approximation) = Angular velocity 2 × Angular velocity ratio obtained by searching from the DB

By applying the velocity 1 (approximation) and angular velocity 1 (approximation) calculated from the above equation to the above equation 100, the self-position 1 (approximation) can be calculated. Therefore, with this method, even if the self-position 1 cannot be obtained due to an error, a self-position (corrected self-position) with the same accuracy as the self-position 1 can be restored from the velocity and angular velocity at the self-position 2. Note that the self-position 2 indicates the position and orientation when an error occurs when the robot 300 acquires the self-position 1.

The hardware configuration of the robot 300 and robot control system according to an embodiment of the present invention will be described using Figures 3A to 3C. Figure 3A is a block diagram showing an example of the hardware of the robot control system. Figure 3B is a perspective view of the robot 300, and Figure 3C is a side view of the robot 300.

The robot control system of the present invention is composed of a robot 300 and a server 320. Note that while this embodiment will be described using a simple example of one robot and one server, it can also be applied to a large-scale robot control system with multiple robots 300 and multiple servers 1.

Robot 300 is an automatic guided vehicle (AGV) designed to transport cargo in a logistics warehouse. Robot 300 is composed of a stereo camera 301, wheels 303 (drive wheels), an encoder 302, a lifter 304 for lifting and lowering cargo, a wheel motor 306 for driving the wheels, a battery 305, PC1 (307), a power source 308, and driven wheels (not shown) for supporting the load.

PC1 (307) is a well-known computer configured with a CPU, memory, a storage device such as an SSD (Solid State Drive), Wi-Fi (or a wireless communication unit), etc. The robot 300 of this embodiment is a differential two-wheel robot that moves forward and rotates using two wheels 303, and includes two sets of wheels 303, encoders 302, and wheel motors 306.

The two wheels 303 are arranged parallel to the direction of travel, and two encoders 302 are attached to each. A battery 305 supplies power to the wheel motor 306, and a power supply 308 supplies power to PC1 (307).

The server 320 is composed of a PC2 (321) and a power supply 322. The PC2 (321) is a well-known computer composed of a CPU, memory, an SSD, Wi-Fi, and the like.
The robot 300 and the server 320 are connected via a wireless LAN, and are capable of communicating using a predetermined protocol (for example, TCP).

The robot 300 in this embodiment is equipped with two types of sensors for recognizing its own position. The first sensor is a stereo camera 301. In this embodiment, the server 320 calculates its own position (self-position 1) from the images captured by the stereo camera 301 using the Visual SLAM function described in Non-Patent Document 1. The Visual SLAM function will be described later.

The second sensor is a pair of encoders 302. The pair of encoders 302 provide the rotation angles of the left and right wheels 303, and the server 320 calculates the speed and angular velocity from the encoder values. The method for calculating self-position 1 will be described later.

Next, the software configuration of the robot 300 and robot control system in Example 1 of the present invention will be described using Figure 4. Figure 4 is a block diagram showing an example of software for the robot control system.

In the figure, block 400 shows the software configuration running on robot 300, and block 420 shows the software configuration running on server 320. These software run on a Linux OS (not shown).

First, we will explain the method for calculating self-position 1. The image acquisition and distribution unit 401 is a program that acquires stereo images from the stereo camera 301 and distributes them to Visual SLAM 422.

Visual SLAM 422 is a program that calculates self-position 1 using image feature points extracted from stereo images, and functions as a self-position 1 calculation unit. A representative Visual SLAM program is "OpenVSLAM: A Versatile Visual SLAM Framework" from the aforementioned Non-Patent Document 1.

This program has two modes: one in which map creation and self-location estimation are performed simultaneously, and another in which a map created in the first mode is read and self-location estimation is performed by comparing the map with camera images. In this example, Visual SLAM 422 is intended to operate in the latter mode.

Specifically, Visual SLAM 422 reads a map 421 created in advance and compares the map 421 with the stereo images obtained from the stereo camera 301 to estimate the vehicle's own position 1.

Visual SLAM 422 outputs a self-position 1 and an error signal as a result of program processing. Self-position 1 is represented by (x, y, θ). x and y indicate the coordinates of the robot 300 on the map 421, and θ is the orientation of the robot 300 (Equation 100 in Figure 1A).

An error signal is output from Visual SLAM 422 if matching between map 421 and stereo camera images fails. That is, Visual SLAM 422 sets the error signal to "1" if self-position 1 cannot be obtained, and sets the error signal to "0" if self-position 1 is obtained. Self-position 1 and the error signal are delivered to the self-position correction unit 426, for example, at a frequency of 30 times per second.

Next, we will explain the method for calculating velocity 2 and angular velocity 2 from encoder 302. The encoder value acquisition and distribution unit 402 is a program that reads the encoder value representing the rotation angle of the wheel 303 from encoder 302 and distributes it to the velocity 2 and angular velocity 2 calculation unit 423.

The velocity 2 and angular velocity 2 calculation unit 423 acquires and distributes the encoder values of each of the left and right wheels 303 approximately 100 times per second. The velocity 2 and angular velocity 2 calculation unit 423 calculates the travel distance per unit time of each wheel 303 from the increase or decrease in the encoder value per unit time.

The distance traveled when wheel 303 rotates once (360 degrees) can be calculated by multiplying the diameter of wheel 303 by the ratio of the circumference of the wheel to the circumference of the circle. Therefore, the distance traveled by robot 300 can be calculated by calculating the rotation angle of wheel 303 from the output of encoder 302.

The velocity 2 and angular velocity 2 calculation unit 423 calculates the velocity 2 (ν2) and angular velocity 2 (ω2) of the robot 300 using equation 200 in Figure 2, assuming that the travel distance of the left wheel is dl and the travel distance of the right wheel is dr, and transmits the velocity 2 (ν2) and angular velocity 2 (ω2) to the self-position correction unit 426.

The self-position correction unit 426 is a program that receives as input the self-position 1 and error signal delivered from the visual SLAM 422, and the velocity 2 and angular velocity 2 delivered from the velocity 2 and angular velocity 2 calculation unit 423, and outputs a corrected self-position (estimated position and attitude) using correction information 702 stored in the database (DB) 425 when the self-position 1 cannot be obtained due to an error. Details of this program will be described later.

On the other hand, the path following program 427 is a program that uses the corrected self-position and path information to make the robot 300 travel along a preset path. The path following program 427 generates information for driving the wheel motor 306, and drives the wheel motor 306 via the motor drive information distribution unit 424 and the motor drive unit 403. Note that the path following program 427, motor drive information distribution unit 424, and motor drive unit 403 can use well-known or publicly known technology, so detailed explanations of their functions will be omitted. This concludes the explanation of the software blocks.

Next, the processing performed by the self-position correction unit 426 will be described in detail using Figure 5. Figure 5 is a flowchart showing an example of the self-position correction processing. This processing consists of three processes: a main loop processing (500), a correction information registration processing (510), and a self-position correction processing (520).

The correction information registration process 510 is a process for registering correction information 702 (see Figure 7C) for restoring a self-position (corrected self-position) with the same accuracy as the first self-position (self-position 1) from the velocity and angular velocity when the error signal becomes "1." The self-position correction process 520 is a process for correcting the self-position using the registered correction information.

First, the main loop processing (start 500 to end 508) will be explained. The program starts at start 500, and in the subsequent processing 501, the self-position correction unit 426 initializes the variable "previous self-position" to (0, 0, 0). "Previous self-position" is a variable that represents the position and orientation (x, y, θ) of the robot 300, and stores the position and orientation of the robot 300 one step before.

In the next step 502, the self-position correction unit 426 waits until it receives self-position 1 and an error signal. The self-position 1 and error signal are transmitted, for example, from Visual SLAM 422 at a frequency of 30 times per second.

Self-position 1 is the position and orientation (x, y, θ) of robot 300 obtained from Visual SLAM 422. The error signal indicates that no error has occurred when it is "0," and indicates that an invalid value has been set for Self-position 1 when it is "1" due to an error.

After receiving self-position 1 and the error signal, the self-position correction unit 426 proceeds to process 503. In process 503, the self-position correction unit 426 acquires velocity 2 and angular velocity 2 transmitted from the velocity 2 and angular velocity 2 calculation unit 423. As described above, velocity 2 and angular velocity 2 are velocity 2 and angular velocity 2 of the robot 300 calculated from the rotation information of the wheels 303.

In process 504, the conditions are determined: the error signal is "0" (self-position 1 was obtained without an error) and the DB registration flag (described below) is "1" (registration is performed in correction information 702 of database 425). If the determination result is Yes, correction information registration process 510 is executed.

The DB registration flag is assumed to be set to "0" or "1" by the user in the program startup argument, but it may also be set to "0" or "1" during program execution according to specified conditions. Details of the correction information registration process 510 will be described later.

If the result of the determination in process 504 is No, then in process 505, a determination is made as to whether or not to execute self-position correction process 520. Specifically, if the error signal is "1" (an error has occurred and self-position 1 cannot be obtained), self-position correction process 520 is executed. Details of self-position correction process 520 will be described later.

In the next process 506, the self-position correction unit 426 updates the previous self-position so that self-position 1 can be used as the previous self-position in the next step. The self-position correction unit 426 also outputs self-position 1 as the corrected self-position. The corrected self-position is the position and orientation (x, y, θ) of the robot 300 after correction in the self-position correction process 520, and if the error signal = "0", the self-position is 1 without correction.

In the next step 507, a program termination determination is made. If the result is Yes, the program proceeds to end step 508 and terminates. If the result is No, the program returns to step 502 and continues. The program termination determination determines whether the self-position correction unit 426 has satisfied the preset termination conditions.

In this embodiment, the series of processes from process 502 to process 507 is referred to as one step. One step is executed, for example, at a 1/30 second interval, which is the reception cycle for self-position 1. This completes the main loop processing.

Next, we will explain the correction information registration process 510. This process is executed if no errors occur in process 504, the correct self-position 1 is obtained, and the DB registration flag is set to "1" (valid).

In process 511, the self-position correction unit 426 calculates velocity 1 (ν1) and angular velocity 1 (ω1) at each time t from the previous self-position and self-position 1 using the following equation 201.

Specifically, let previous self-position = (x(t-1), y(t-1), θ(t-1)), self-position 1 = (x(t), y(t), θ(t)), and substitute Δt = 1/30th of a second into the above formula 201 to calculate velocity 1 (ν1) and angular velocity 1 (ν1).

In the next process 512, the self-position correction unit 426 calculates the velocity ratio from velocity 1 and velocity 2, and the angular velocity ratio from angular velocity 1 and angular velocity 2. Furthermore, in the next process 513, the self-position correction unit 426 links the velocity ratio and angular velocity ratio to the previous self-position (x(t-1), y(t-1), θ(t-1)), velocity 2 (ν2), and angular velocity 2 (ω2), and registers them in the correction information 702 of the DB 425. This completes the correction information registration process 510.

Next, we will explain the self-position correction process 520. The self-position correction process 520 is executed if an error occurs in process 505 and self-position 1 cannot be obtained.

In process 521, the self-position correction unit 426 uses the previous self-position (x, y, θ), velocity 2 (ν2), and angular velocity 2 (ω2) as search keys to search for the most matching velocity ratio and angular velocity ratio from the correction information 702 in DB 425. Details of this process will be described later.

If a velocity ratio and angular velocity ratio are obtained as search results, the process branches to Yes in the judgment of process 522. If a velocity ratio and angular velocity ratio are not obtained, the process branches to No, and in process 532, the velocity ratio = 1.0 and angular velocity ratio = 1.0 are set. In the next process 524, the estimated values of velocity 1 and angular velocity 1 that would have been obtained if no error had occurred are calculated using the following equation 202.

Velocity 1 = Velocity 2 x Velocity ratio Angular velocity 1 = Angular velocity 2 x Angular velocity ratio ... (202)

In the next process 525, the self-position correction unit 426 calculates self-position 1 (estimated position and estimated attitude) from the previous self-position, the estimated value of velocity 1, and the estimated value of angular velocity 1 using equation 100 in Figure 1. Specifically, self-position 1 (x(t), y(t), θ(t)) is calculated using (x(t-1), y(t-1), θ(t-1)) = previous self-position, ν(t) = velocity 1 (ν1), ω(t) = angular velocity 1 (ω1), and Δt = 1/30th of a second.

When process 525 is complete, proceed to process 506. This completes self-position correction process 520.

If the error continues to occur, the self-position correction unit 426 searches for correction information 702 using the self-position 1 (estimated position and estimated attitude), velocity 2 (ν2) at the estimated position, and angular velocity 2 (ω2) at the estimated attitude, and obtains the velocity ratio and angular velocity ratio from the search results as described above.

Then, the self-position correction unit 426 calculates an estimate of velocity 1 (ν1) at the estimated position and an estimate of angular velocity 1 (ω1) at the estimated attitude from the acquired velocity ratio and angular velocity ratio, velocity 2 (ν2) at the estimated position, and angular velocity 1 (ω2) at the estimated attitude, and calculates self-position 1 based on the calculated estimate of velocity 1 (ν1) at the estimated position and the estimated value of angular velocity 1 (ω1) at the estimated attitude.

Next, the details of the DB search process of process 521 in Figure 5 will be explained using Figure 6. This process is implemented, for example, as a function, and argument 600 in the figure is the argument of that function.

The position and orientation (x, y, θ) of the robot 300 are set as the search keys ix, iy, iθ, velocity 2 (ν2) and angular velocity 2 (ω2) are set as inv2 and iω2, the search ranges s1, s2, s3, s4, and s5 are set to the search ranges of ix, iy, iθ, inv2, and iω2, and mode is set to mode = 0 to return the velocity and angular velocity with the highest degree of match, or modo = 1 to return the average of the matched velocity and angular velocity, and this function is executed with this argument set.

In process 601, the self-position correction unit 426 searches for the velocity ratio and angular velocity ratio near the current position (x, y) of the robot 300 from the velocity ratios and angular velocity ratios stored in the correction information 702 of the DB 425 using the following conditional expression 203.

(ix-s1<x<ix+s1) and (iy-s2<y<iy+s2)
...(203)

Note that s1 and s2 are preset constants for searching for data in the vicinity (predetermined range) of the positions ix and iy that serve as search keys.
If the self-position correction unit 426 finds even one velocity ratio and angular velocity ratio, it branches to Yes in process 602, and if not, it branches to No.

In process 603, the self-position correction unit 426 further narrows down the search from the correction information 702 found in the above search to find records that satisfy the following conditional expression 204 for the orientation, velocity ratio, and angular velocity ratio.

(iθ−s3 < θ < iθ+s3) and
(iν2-s4 < ν2 < iν2+s4) and
(iω2-s5 < ω2 < iω2+s5) ......(204)

Note that s3, s4, and s5 are preset constants used to search for data in the vicinity (predetermined range) of the direction θ, velocity 2ν2, and angular velocity 2ω2, which serve as the second search key.

Because a DB search using multiple search keys slows down the search speed, in this embodiment, a program (self-position correction unit 426) performs a refined search on the results of a rough search of correction information 702 in DB 425. If at least one velocity ratio and angular velocity ratio is found as a result of the refined search, processing 604 branches to Yes; if not, processing 604 branches to No.

In process 605, the self-position correction unit 426 branches depending on the search mode (mode). If modo = 0, process 608 returns the velocity ratio and angular velocity ratio with the highest degree of match.

Specifically, the self-position correction unit 426 selects and returns the velocity ratio and angular velocity ratio that minimize the Euclidean distance between each variable from the results obtained by the narrowed search.

If modo = 1, in process 607, the self-position correction unit 426 calculates and returns the average speed ratio and angular velocity ratio from the results obtained by the narrowed search. If processes 602 and 604 branch to No, the self-position correction unit 426 returns "search failed." This concludes the details of the DB search process in process 521.

Next, examples of robot movement and the self-position correction unit 426 will be described using Figures 7A and 7B. Figure 7A is a diagram showing an example of correction information when the robot 300 moves straight ahead, and Figure 7B is a diagram showing an example of correction information when the robot 300 turns right.

In Figure 7A, trajectory 700 shows an example of the robot moving straight in the direction of the arrow, and in Figure 7B, trajectory 701 shows an example of the robot 300 turning right.

When registering correction information 702, the robot 300 is moved in the direction of the arrow along trajectories 700 and 701 in the figure, and correction information (previous self-position, velocity 2, angular velocity 2, velocity ratio, angular velocity ratio) 702 is registered in DB 425.

Correction information 702 is information that registers position information (P100 to P120 in Figure 7A, P200 to P260 in Figure 7B) obtained from robot 300 while it is moving, and the moving conditions at each position.

Figure 7C shows an example of correction information 702. The correction information 702 includes a legend 7021, a position 7022, an orientation 7023, a velocity 7024, a velocity ratio 7025, an angular velocity ratio 7026, and a remark 7027 in one record.

The legend 7021 stores the identifier set by the server 320. The position 7022 stores the position (x, y) of the robot 300, and the orientation 7023 stores the orientation θ of the robot 300.

Velocity 7024 stores a search key consisting of velocity 2 (ν2) and angular velocity 2 (ω2) obtained from the encoder 302. Velocity ratio 7025 stores velocity 2 calculated using equation 200 above. Angular velocity ratio 7026 stores angular velocity ω2 calculated using equation 200 above. Remarks 7027 stores items specified by the server 320.

The correction information 702 stores a search key consisting of the robot 300's position 7022, orientation 7023, velocity 2 (ν2) and angular velocity 2 (ω2) obtained from the encoder 302, and associates the velocity ratio 7025 and angular velocity ratio 7026 with each other.

For the sake of convenience, in the illustrated example, only a few items of correction information 702 are registered; however, in reality, correction information 702 is registered 30 times per second while the robot 300 is traveling along a predetermined path, which corresponds to approximately every 2-3 cm of the robot's travel distance.

The correction information (P110, P220, P230, P240) shown in Figures 7A and 7B is correction information for approximately the same location, but the orientation of robot 300, velocity 2 (ν2), and angular velocity 2 (ω2) are different, and correction information (velocity ratio, angular velocity ratio) is registered according to each state. Roughly speaking, P110 in the figures indicates correction information when moving straight, P210 in the figures indicates correction information when decelerating, P220 in the figures indicates correction information when turning right, and P240 in the figures indicates correction information when turning right while decelerating.

It is believed that the robot 300 in the logistics warehouse travels along a predetermined route. Therefore, by collecting correction information 702 while the robot travels along the predetermined route and registering it in DB 425, correction information 702 near the robot 300 can be reliably obtained when an error occurs.

Figure 8 shows an example of correction information for each type of movement, such as when turning right. It is considered that correction information 702 according to the type of movement of the robot 300 is required when going straight or turning right.

In the figure, trajectories 800L and 800R are the trajectories of the left and right wheels 303 when the robot 300 moves straight. trajectories 801L and 801R are the trajectories of the left and right wheels 303 when the robot 300 turns right.

Area 803 is a location where the road surface has a low coefficient of friction, making it easy for wheels 303 to slip. In the example shown, robot 300 does not pass over area 803 when traveling straight, but does pass over it when turning right.

When turning right, wheels 303 are more likely to slip, so it is thought that the angular velocity ratio (ν1/ν2) will be somewhat smaller than 1. On the other hand, when traveling straight, it is thought that an angular velocity ratio very close to 1 will be obtained. Therefore, by registering correction information 702 according to the type of movement of robot 300 (straight traveling, right turning, etc.) at point 802, it is thought that it will be possible to correct the robot's own position with greater accuracy depending on the type of movement (straight traveling, right turning, etc.) and the road surface conditions, even in the same location.

As a result of the above, in Example 1 of the present invention, even if the first self-location cannot be obtained due to an error, the first self-location can be calculated by correcting the second self-location using the correction information 702. In this case, even if the robot is in the same location, the self-location can be calculated with high accuracy according to the position, posture, and type of movement (direction, straight forward, rotation) of the robot 300.

In the above-described first embodiment, an example was shown in which a stereo camera 301 was used as the first sensor and odometry was used as the second sensor, but this is not limited to this and any sensor capable of detecting the position of the robot 300 with high accuracy may be used. For example, a LiDAR (Light Detection And Ranging) sensor may be used as the first sensor, and a GNSS (Global Navigation Satellite System) may be used as the second sensor if the location where the robot 300 is operated is a location where communication with satellites is possible.

In addition, in the above-mentioned Example 1, an example was shown in which the server 320 calculated the self-position 1 and corrected the self-position, and then drove the wheel motor 306 at the corrected self-position 1, but this is not limited to this. Although not shown in the figures, the map 421, Visual SLAM 422, velocity 2 and angular velocity 2 calculation unit 423, database 425 (correction information 702), self-position correction unit 426, path following program 427, and motor drive information distribution unit 424 may be operated by the robot 300 to correct the self-position.

Furthermore, the above-described first embodiment can be applied to a transport system in which the robot 300 transports objects such as luggage and merchandise.

Example 2 describes the differences from Example 1.

(1) The following items may be added as correction information 702 to be registered in DB425.

The addition of the "Date and Time" item makes it possible to selectively search for the latest correction information 702. It also makes it possible to correct for changes in the friction coefficient over time due to road surface wear and deterioration of the robot over time.

The addition of the "Acceleration" and "Angular Acceleration" items makes it possible to correct the robot 300's position according to more detailed types of movement.

The addition of the "Load Weight" item makes it possible to correct the position when the circumference of the wheels 303 changes due to the weight of the robot 300. However, the "Load Weight" item is applicable when the wheels 303 are elastically deformed. Also, if the wheels 303 are pneumatic tires, the addition of the "Wheel Air Pressure" item makes it possible to correct when the circumference of the wheels 303 changes depending on the air pressure of the wheels 303. The addition of the "Wheel Circumference" item makes it possible to correct when the circumference of the wheels 303 changes due to wear on the wheels 303.

The addition of the "Wheel Torque" item allows for corrections based on the road surface's friction coefficient. The addition of the "Temperature" and "Humidity" items allows for corrections based on the wetness of the road surface.

The addition of the "Robot Unique ID" item makes it possible to make corrections according to the habits and characteristics of each individual robot 300.

(2) In process 504 of Figure 5 in Example 1, it was stated that "during program execution, the DB registration flags for the velocity ratio and angular velocity ratio may be set to "0" or "1" according to certain conditions." The following describes the conditions for dynamically setting the DB registration flags to "1."

In logistics warehouses, robots 300 may be stopped outside of working hours or during breaks. During such times when robots 300 are not performing transport work, one possible solution is to set the DB registration flag to 1, run the robot 300 along the route, and update the correction information 702 in DB 425.

Visual SLAM 422 loads a pre-created map 421 and compares the map 421 with the stereo images obtained from the stereo camera 301 to estimate the vehicle's position.

Specifically, Visual SLAM 422 extracts image feature points from stereo images and compares them with feature points (map feature points) in map 421 to estimate its own position. In doing so, it compares the number of image feature points obtained from stereo camera 301 with the number of feature points in map 421, and if there is a difference in the number of feature points of a certain number or more, it is determined that a change has occurred in the environment, and one possible solution is to update the correction information 702 in DB 425.

On the other hand, it is possible to keep the DB registration flag set to 1 and continuously register correction information 702 such as speed ratio and acceleration ratio in DB 425.

(3) Regarding the DB registration flag in process 513, in addition to the velocity ratio and angular velocity ratio, it is also possible to store velocity 1 (ν1), angular velocity 1 (ω1), velocity 2 (ν2), and angular velocity 2 (ω2) that can calculate these. The database may also be in the form of a key-value store or JSON document structure. Furthermore, in addition to the rule-based method shown in Example 1, the velocity ratio and angular velocity ratio may be calculated using machine learning or other methods by inputting position information, orientation, and velocity (x, y, θ, ν2, ω2).

(4) This section describes the action of the robot 300 intended to notify the user when performing the self-position correction process 520 shown in Figure 5 of Example 1. This action is executed during the self-position correction process 520, such as immediately after process 525.

Specifically, when the self-position correction process 520 is executed, if the self-position correction process 520 continues beyond a certain distance, the robot 300 is stopped. Alternatively, while the self-position correction process 520 is continuously executed, the robot 300 is made to travel at a slower speed than normal. Alternatively, the robot 300 is made to travel while its body lamps (tail lamps, blinkers 309, etc.) are turned on or flashing, or the flashing pattern is made to speed up or slow down. Alternatively, the robot 300 may be made to travel while sounding the horn 310 or playing music, etc.

In this way, in sections where the self-position correction process 520 occurs frequently, the robot 300 can emit light and sound to alert the user to the possibility that some kind of abnormality has occurred.

<Conclusion>
As described above, the robot control system of each of the above embodiments can be configured as follows.

(1) A robot (300) control system in which a server (320) having a processor and memory estimates the position of a mobile robot (300), the mobile robot (300) having a first sensor unit (stereo camera 301) that acquires first sensing information (stereo images) from the traveling state of the mobile robot (300), and a second sensor unit (encoder 302) that has a sensor different from the first sensor unit (301) and acquires second sensing information (rotation angle of wheels 303) from the traveling state of the mobile robot (300), the server (320) acquires first traveling state information including a first position and a first attitude (θ) at the first position obtained from the first sensing information (stereo images), and a second velocity (ν2) and a second angular velocity (ω2) of the mobile robot (300) obtained from the second sensing information (rotation angle of wheels 303). and second driving state information including information on the first position and the second driving state information, and a calculation unit (self-position correction unit 426) capable of calculating an estimated position (corrected self-position) of the mobile robot (300) based on the information on the first driving state and the information on the second driving state. The calculation unit (426) calculates the first position and the first attitude (θ) based on first sensing information (stereo images) acquired from the first sensor unit (301), and, if an error occurs during the calculation of the first position, calculates the estimated position (corrected self-position) of the mobile robot (300) based on second driving state information (velocity 2, angular velocity 2) acquired from the second sensing information (rotation angle of the wheels 303) and the correspondence information (correction information 702).

With the above configuration, even if an autonomously traveling automated guided vehicle (mobile robot) is unable to obtain its own position 1 (first position) calculated from the stereo images (first sensing information) of the stereo camera 301 (first sensor unit) due to an error, it is possible to restore its own position (corrected own position) with the same accuracy as the first position from the second traveling state (velocity 2, angular velocity 2) calculated from information from the encoder 302 (second sensor) and correction information 702 (correspondence information).

(2) The robot control system described in (1) above, wherein the correspondence information (correction information 702) includes the velocity ratio (ν1/ν2) between the first velocity (ν1) and the second velocity (ν2) at the first position (self-position 1), and the angular velocity ratio (ω1/ω2) between the first angular velocity (ω1) and the second angular velocity (ω2) at the first attitude (θ).

With the above configuration, velocity 1 (ν1) can be calculated from velocity 2 (ν2) obtained from encoder 302 and the velocity ratio (ν1/ν2) recorded in correction information 702, angular velocity 1 (ω1) can be calculated from second angular velocity (ω2) obtained from encoder 302 and the angular velocity ratio (ω1/ω2) recorded in correction information 702, and corrected self-position (self-position 1) can be calculated from velocity 1 (ν1) and angular velocity 1 (ω1). As a result, even if self-position 1 cannot be obtained from stereo images due to an error, corrected self-position can be calculated based on velocity 2 and angular velocity 2 obtained from encoder 302, and self-position 1 (corrected self-position) with the same accuracy as self-position 1 can be restored.

(3) The robot control system described in (2) above, wherein, when there is no error and the preset registration information (DB registration flag) is valid ("1"), the calculation unit (self-position correction unit 426) acquires the first velocity (ν1) and the first angular velocity (ω1), calculates the velocity ratio (ν1/ν2) and the angular velocity ratio (ω1/ω2) from the second velocity (ν2) and the second angular velocity (ω2), and registers the second velocity (ν2), the second angular velocity (ω2), the velocity ratio (ν1/ν2), and the angular velocity ratio (ω1/ω2) in the correspondence information (correction information 702) in association with the first position (self-position 1) and the first attitude (θ).

With the above configuration, if there is no error in the calculation of self-position 1 and the DB registration flag is enabled, the self-position correction unit 426 acquires velocity 1 and angular velocity 1, and associates the velocity ratio (ν1/ν2) and angular velocity ratio (ω1/ω2) from velocity 2 (ν2) and angular velocity 2 (ω2) with self-position 1 and the first attitude (θ), and registers these in the correspondence information (correction information 702), which can be used to restore self-position 1 in the event of an error. Note that velocity 1 and angular velocity 1 may also be calculated from the difference between the previous and current values of self-position 1.

(4) The robot control system described in (2) above, wherein, when the error occurs, the calculation unit (self-position correction unit 426) searches the correspondence information (correction information 702) using information on the first position (self-position 1) and the first orientation (θ) of the mobile robot before the error occurred, the second velocity (ν2), and the second angular velocity (ω2), obtains the velocity ratio (ν1/ν2) and the angular velocity ratio (ω1/ω2) from the search results, and calculates the position and orientation of the mobile robot as an estimated position and estimated orientation based on the estimated value of the first velocity (ν1) and the estimated value of the first angular velocity (ω1) from the obtained velocity ratio (ν1/ν2) and angular velocity ratio (ω1/ω2), the second velocity (ν2), and the second angular velocity (ω2).

With the above configuration, if an error occurs in the calculation of self-position 1, the self-position correction unit 426 searches correction information 702 for velocity 2 and angular velocity 2 calculated from self-position 1 and information from encoder 302, and obtains the velocity ratio and angular velocity ratio from the search results. The self-position correction unit 426 then calculates velocity 1 from the velocity ratio and velocity 2, calculates angular velocity 1 from the angular velocity ratio and angular velocity 2, and can calculate self-position 1 and attitude from self-position 1 and the calculated velocity 1 and angular velocity 1.

(5) In the robot control system described in (4) above, when the error occurs, the calculation unit (self-position correction unit 426) searches the correspondence information (correction information 702) using the estimated position and estimated attitude calculated by the calculation unit (426), the second velocity (ν2) at the estimated position, and the second angular velocity (ω2) at the estimated attitude, obtains the velocity ratio (ν1/ν2) and angular velocity ratio (ω1/ω2) from the search results, and a robot control system that calculates an estimate of the first velocity (ν1) at the estimated position and an estimate of the first angular velocity (ω1) at the estimated posture from (ω1/ω2), the second velocity (ν2) at the estimated position, and the second angular velocity (ω2) at the estimated posture, and calculates the position and posture of the mobile robot based on the calculated estimate of the first velocity (ν1) at the estimated position and the estimated value of the first angular velocity (ω1) at the estimated posture.

With the above configuration, if an error continues to occur in the calculation of self-position 1, the self-position correction unit 426 can continue to estimate the position and attitude of the mobile robot by using the calculated estimated position and attitude, even while the error continues.

(6) The robot control system described in (4) above, wherein the calculation unit (self-position correction unit 426), when searching the correspondence information (correction information 702), searches using a first key representing the first position (self-position 1), and further searches the search results using a second key representing a first attitude (θ), a second velocity (ν2), and a second angular velocity (ω2).

With the above configuration, the self-position correction unit 426 can narrow down the results of a rough search using the first key using the second search key, thereby preventing the DB search speed from slowing down when using multiple search keys.

(7) The robot control system described in (2) above, wherein the correspondence information (correction information 702) includes the first position (self-position 1), the first attitude (θ), the velocity ratio (ν1/ν2), and the angular velocity ratio (ω1/ω2).

With the above configuration, the correspondence information (correction information 702) can include the first position (self-position 1), the first attitude (θ), the velocity ratio (ν1/ν2), and the angular velocity ratio (ω1/ω2).

(8) The robot control system described in (7) above, wherein the correspondence information (correction information 702) includes, in addition to the first position (self-position 1), the first attitude (θ), the velocity ratio (ν1/ν2), and the angular velocity ratio (ω1/ω2), at least one of the following: date and time, acceleration, angular acceleration, weight of the payload, wheel air pressure, wheel circumference, wheel torque, temperature, humidity, and a unique ID of the mobile robot (300).

With the above configuration, adding "date and time" allows for selective retrieval of the latest correction information 702. Adding "acceleration" and "angular acceleration" allows for more precise correction of the robot 300's position according to the type of movement. Adding "load weight" allows for position correction when the circumference of the wheels 303 changes due to the weight of the robot 300. If the wheels 303 are pneumatic tires, adding "wheel air pressure" allows for correction when the circumference of the wheels 303 changes depending on the air pressure of the wheels 303. Adding "wheel circumference" allows for correction when the circumference of the wheels 303 changes due to wear of the wheels 303. Adding "wheel torque" allows for correction according to the friction coefficient of the road surface. Adding "temperature" and "humidity" allows for correction according to the wetness of the road surface. Adding "robot-specific ID" allows for correction according to the habits and characteristics of each individual robot 300.

(9) The robot control system described in (1) above, wherein, when the first position (self-position 1) cannot be acquired due to the error, the calculation unit (self-position correction unit 426) commands the mobile robot (300) to at least one of blinking a lamp (blinker 309) provided on the mobile robot (300), outputting a sound from a horn (310) provided on the mobile robot (300), and slowing down or stopping the mobile robot (300).

With the above configuration, if an error occurs in the calculation of self-position 1, the robot 300's blinker 309 can be flashed, the horn 310 can be sounded, and the wheel motor 306 can be slowed down or stopped to notify the user that an abnormality has occurred in the calculation of self-position 1.

The present invention is not limited to the above-described embodiments, and includes various modifications. For example, the above-described embodiments have been described in detail to clearly explain the present invention, and are not necessarily limited to including all of the configurations described. Furthermore, it is possible to replace part of the configuration of one embodiment with the configuration of another embodiment, or to add the configuration of another embodiment to the configuration of one embodiment. Furthermore, it is possible to add, delete, or replace part of the configuration of each embodiment with other configurations, either alone or in combination.

Furthermore, the above-mentioned configurations, functions, processing units, and processing means may be implemented in part or in whole in hardware, for example by designing them as integrated circuits. Furthermore, the above-mentioned configurations and functions may be implemented in software by a processor interpreting and executing programs that implement each function. Information such as programs, tables, and files that implement each function can be stored in memory, a recording device such as a hard disk or SSD, or a recording medium such as an IC card, SD card, or DVD.

Furthermore, the control and information lines shown are those deemed necessary for explanation, and do not necessarily represent all control and information lines on the product. In reality, it is safe to assume that almost all components are interconnected.

300 Robot 302 Stereo camera 302 Encoder 303 Wheel 421 Map 422 Visual SLAM
423 Velocity 2, angular velocity 2 calculation unit 425 Database 426 Self-position correction unit 702 Correction information

Claims (11)

A robot control system in which a server having a processor and a memory estimates a position of a mobile robot,
The mobile robot is
a first sensor unit that acquires first sensing information from a running state of the mobile robot;
a second sensor unit having a sensor different from the first sensor unit and configured to acquire second sensing information from the traveling state of the mobile robot;
The server
a storage unit that stores correspondence information indicating a correspondence between first running state information including a first position and a first orientation at the first position obtained from the first sensing information, and second running state information including a second velocity and a second angular velocity of the mobile robot obtained from the second sensing information, in association with the first position;
a calculation unit capable of calculating an estimated position of the mobile robot based on the information on the first running state and the information on the second running state,
The calculation unit
A robot control system characterized in that the first position and the first attitude are calculated based on first sensing information acquired from the first sensor unit, and if an error occurs when calculating the first position, an estimated position of the mobile robot is calculated based on second running state information acquired from the second sensing information and the correspondence information. 2. The robot control system of claim 1,
The correspondence information is
a velocity ratio between a first velocity and the second velocity at the first position, and an angular velocity ratio between a first angular velocity and the second angular velocity at the first posture. 3. The robot control system according to claim 2,
The calculation unit
and, if there is no error and the preset registration information is valid, acquiring the first velocity and the first angular velocity, calculating the velocity ratio from the first velocity and the second velocity, calculating the angular velocity ratio from the first angular velocity and the second angular velocity, and registering the second velocity, the second angular velocity, the velocity ratio, and the angular velocity ratio in the correspondence information in association with the first position and the first attitude. 3. The robot control system according to claim 2,
The calculation unit
and when the error occurs, searching for the correspondence information using information on the first position and first orientation of the mobile robot before the error occurred, the second velocity, and the second angular velocity, obtaining a velocity ratio and an angular velocity ratio from the search results, calculating an estimated value of the first velocity and an estimated value of the first angular velocity from the obtained velocity ratio and angular velocity ratio, the second velocity, and the second angular velocity, and calculating an estimated position and orientation of the mobile robot based on the calculated estimated value of the first velocity and the first estimated value of the first angular velocity. 5. The robot control system according to claim 4,
The calculation unit
and when the error occurs, searching the correspondence information using the estimated position and estimated attitude calculated by the calculation unit, the second velocity at the estimated position, and the second angular velocity at the estimated attitude, obtaining a velocity ratio and an angular velocity ratio from the search result, calculating an estimate of the first velocity at the estimated position and an estimate of the first angular velocity at the estimated attitude from the obtained velocity ratio and angular velocity ratio, the second velocity at the estimated position, and the second angular velocity at the estimated attitude, and calculating the position and attitude of the mobile robot based on the calculated estimate of the first velocity at the estimated position and the calculated estimate of the first angular velocity at the estimated attitude. 5. The robot control system according to claim 4,
The calculation unit
a robot control system characterized in that, when searching for the correspondence information, a search is performed using a first key representing the first position, and the search results are then searched using a second key representing the first attitude, the second velocity, and the second angular velocity. 3. The robot control system according to claim 2,
The correspondence information is
A robot control system comprising the first position, the first attitude, the velocity ratio, and the angular velocity ratio. 8. The robot control system of claim 7,
The correspondence information is
A robot control system characterized in that, in addition to the first position, the first attitude, the velocity ratio, and the angular velocity ratio, the information includes at least one of date and time, acceleration, angular acceleration, weight of a payload, wheel air pressure, wheel circumference, wheel torque, temperature, humidity, and a mobile robot unique ID. 2. The robot control system of claim 1,
The calculation unit
A robot control system characterized in that, when the first position cannot be obtained due to the error, the system instructs the mobile robot to do at least one of flashing a lamp provided on the mobile robot, outputting a sound from a horn provided on the mobile robot, and slowing down and stopping the mobile robot. a server having a processor and a memory;
a mobile robot connected to the server and configured to transport an object,
The mobile robot is
a first sensor unit that acquires first sensing information from a running state of the mobile robot;
a second sensor unit having a sensor different from the first sensor unit and configured to acquire second sensing information from the traveling state of the mobile robot;
The server
a storage unit that stores correspondence information indicating a correspondence between a first position obtained from the first sensing information, first running state information including a first attitude at the first position, and second running state information including a second velocity and a second angular velocity of the mobile robot obtained from the second sensing information, in association with the first position;
a calculation unit capable of calculating an estimated position of the mobile robot based on the information on the first running state and the information on the second running state,
The calculation unit
a conveying system that calculates the first position and the first attitude based on first sensing information acquired from the first sensor unit, and, if an error occurs when calculating the first position, calculates an estimated position of the mobile robot based on second running state information acquired from the second sensing information and the correspondence information. A robot control method in which a server having a processor and a memory estimates a position of a mobile robot, comprising:
a first step in which the server acquires first sensing information detected by a first sensor unit of the mobile robot;
a second step in which the server acquires second sensing information detected by a second sensor unit different from the first sensor unit of the mobile robot;
a third step in which the server stores, in a storage unit, correspondence information indicating a correspondence between first running state information including a first position and a first orientation at the first position obtained from the first sensing information, and second running state information including a second velocity and a second angular velocity of the mobile robot obtained from the second sensing information, in association with the first position;
a fourth step in which the server calculates the first position and the first attitude based on first sensing information acquired from the first sensor unit;
a fifth step in which, when an error occurs in calculating the first position, the server calculates an estimated position of the mobile robot based on second running state information acquired from the second sensing information and the correspondence information;
A robot control method comprising: