Robot motion control method based on artificial intelligence

Abstract

The invention discloses a robot motion control method based on artificial intelligence, and particularly relates to the field of motion control, and the method comprises the steps: firstly collecting visual data and depth data, including front images, continuous frames and point cloud coordinates of a current position, of a robot in a driving process through a high-definition camera and a laser radar; and processing the collected data, and identifying obstacles in the environment and position coordinates of the obstacles. Then, generating a motion path of the robot according to the position and the shape of the obstacle, and planning an optimal path from a starting position to a target position on the discrete grid map by using an A * algorithm; and finally, the motion path is received and executed, the speed and direction of the robot are adjusted through a PID control algorithm, and the robot is made to move along the preset path.

Description

Technical Field

[0001] The present invention relates to the field of motion control technology, and more specifically, to a robot motion control method based on artificial intelligence. Background Art

[0002] With the continuous progress of computer science, sensor technology, etc., robotics technology has developed rapidly. Mobile robots are increasingly used in various fields, such as industrial manufacturing, logistics distribution, medical care, home services, etc.; this requires robots to navigate and move autonomously in complex environments, and global path planning, as the basis and key to robot autonomous movement, is becoming increasingly important. By modeling the environment in advance and planning the global path, the robot can better understand the surrounding environment and find the optimal or suboptimal path from the starting point to the target point, thereby achieving efficient and accurate motion control.

[0003] Robot motion control involves knowledge from multiple disciplines, such as mechanical engineering, electronic engineering, computer science, control theory, etc. The trend of integration between disciplines is constantly strengthening, providing a broader space for the development of robot motion control methods based on global path planning.

[0004] However, traditional robot control methods still have some shortcomings in actual use. For example, early robot motion control methods often rely on simple local perception and response strategies; these methods cannot flexibly respond to the emergence of obstacles and are difficult to adapt to environmental changes when faced with complex and dynamic environments. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a robot motion control method based on artificial intelligence to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] Step A1: Collect visual data around the machine, depth data obtained by the lidar, and robot status data obtained by the sensor;

[0008] Step A2: identifying obstacles and obstacle position coordinates in the environment by analyzing the visual data and the depth data;

[0009] Step A3: Generate a robot motion path based on the identified obstacle position coordinates;

[0010] Step A4: receiving the motion path of the robot, and then controlling the robot to execute motion according to the motion path.

[0011] Preferably, in step A1, a high-definition camera is used to collect visual data, and the visual data includes front image data and continuous frames during the robot's driving process; a laser radar is used to collect the robot's current depth data, and the depth data includes point cloud coordinates and distance data of the robot's current position.

[0012] Preferably, in step A2, a three-dimensional coordinate system is constructed with the robot starting position as the origin, and the collected image data is grayed. The calculation method for converting the color image into the gray image is specifically as follows:

[0013] A(D)=0.2989R+0.5870G+0.1140B, where A(D) represents the converted gray image, R represents the red intensity value, G represents the green intensity value, and B represents the blue intensity value;

[0014] The converted gray image is subjected to a median filter to remove noise;

[0015] The frame difference is calculated in the gray image after denoising. The specific method for calculating the frame difference is:

[0016] B(x, y)=|A(D t+1 (x,y))-A(D t (x, y))|, where B(x, y) represents the frame difference, (x, y) represents the pixel coordinates, and A(D t+1 (x, y)) represents the gray image of the obstacle position at time t+1, A(D t (x, y)) is represented as the gray image of the obstacle position at time t, D t+1 (x, y) represents the image frame collected at time t+1, D t (x, y) represents the image frame collected at time t;

[0017] A threshold is set to judge the frame difference and a binary image is generated. The specific calculation method for generating a binary image is:

[0018] Where C(x, y) represents the pixel value in the binary image, B(x, y) represents the frame difference, T represents the threshold, 1 represents the motion state, and 0 represents the static state;

[0019] Then the frame difference is judged by the threshold value, and the obstacles are classified into moving obstacles and stationary obstacles;

[0020] The specific method for calculating the threshold is:

[0021] T=α+bβ, where T represents the threshold, α represents the average value of the frame difference, and β represents the standard deviation of the frame difference;

[0022] Among them, α represents the average value of all pixel differences, N represents the total number of pixels in the image, and B i (x, y) represents the frame difference of the i-th pixel, and (x, y) represents the pixel coordinates;

[0023] Among them, β represents the standard deviation of all pixel differences, α represents the average value of all pixel differences, N represents the total number of pixels in the image, and B i (x, y) represents the frame difference of the i-th pixel, and (x, y) represents the pixel coordinates;

[0024] The obstacle is wrapped by an n-dimensional ellipsoid to simplify the obstacle. The calculation method of the obstacle surface wrapped by the n-dimensional ellipsoid is as follows:

[0025] where Γ(s) is continuous and has continuous partial derivatives, n is the dimension, and s u Represented as the u-th value of the coordinate in n-dimensional space, Represented as the center of the ellipsoid, p u Expressed as the index of the ellipsoid, r u Expressed as the semi-axis length of the ellipsoid;

[0026] Depend on The four parameters determine an n-dimensional ellipsoid, parameter r u Determines the size of the ellipsoid on each axis, parameter p u Determines the degree of squareness of the ellipsoid, p u The larger the value, the closer the ellipsoid is to a cube. Determine the position of the ellipsoid;

[0027] The specific method for calculating the distance the robot reaches the obstacle is:

[0028] Where L is the distance from the robot to the obstacle, and c is the speed of light, which is 3×10 8 m / s, t 0 It is expressed as the time difference between the emission of the laser pulse and the reception of the reflected pulse;

[0029] The angle calculation method between the robot and the obstacle is as follows:

[0030] Among them, θ represents the angle between the robot and the obstacle, y 2 Represented as the value of the obstacle on the y-axis, y 1 Expressed as the value of the robot on the y-axis, x 2 Represented as the value of the obstacle on the x-axis, x1 It is expressed as the value of the robot on the x-axis;

[0031] A vector is constructed with the robot's position as the starting point and the obstacle's position as the end point. The angle is determined by calculating the angle between the vector and the reference direction. The specific method for determining the angle is as follows:

[0032] The robot's position is F 1 (x 1 ,y 1 ), the position of the obstacle is F 2 (x 2 ,y 2 ), the reference direction is the x-axis, and the calculation method of the vector is as follows:

[0033] in, Represented as a vector, y 2 Represented as the value of the obstacle on the y-axis, y 1 Expressed as the value of the robot on the y-axis, x 2 Represented as the value of the obstacle on the x-axis, x 1 It is expressed as the value of the robot on the x-axis;

[0034] Then the vector The angle with the positive direction of the x-axis is the angle θ between the robot and the obstacle;

[0035] The position coordinates of the obstacle are determined based on the angle and the distance from the robot to the obstacle.

[0036] Preferably, in step A3, the robot's working environment is divided into discrete grids, and the center point of each grid is regarded as a node; the inaccessible grids are marked on the map according to the position and shape of the obstacles; then an open list and a closed list are created, the open list stores the nodes to be processed, and the closed list stores the processed nodes; the starting position and the target position of the robot are marked on the map as the starting point and the end point of the A* algorithm respectively; and the total cost from the starting position to the target position is calculated;

[0037] The total cost is calculated as follows:

[0038] g(d)=h(d)+q(d), where g(d) represents the total cost starting from the initial position, passing through point d and reaching the target position, h(d) represents the actual cost from the starting position to node d, and q(d) represents the estimated cost of the optimal path from node d to the target position.

[0039] Preferably, in step A4, the robot accepts the optimal path for driving, selects PID parameters, and the parameters include proportional gain, integral gain and differential gain; calculates the error according to the PID parameters, and then calculates the control signal according to the error, converts the control signal into the speed reached by the robot tire, and then outputs the calculated robot tire speed to the robot control system to drive the robot to move along the path;

[0040] The speed reached by the left tire is calculated as follows:

[0041] Among them, v l Expressed as the speed reached by the left wheel, v 1 Expressed as linear velocity, L 1 It is represented as the distance from the robot to the target position, φ is represented as the angular velocity, and H is represented as the radius of the tire;

[0042] The speed reached by the right tire is calculated as follows:

[0043] v R Expressed as the speed reached by the right wheel, v 1 Expressed as linear velocity, L 1 It is represented as the distance from the robot to the target position, φ is represented as the angular velocity, and H is represented as the radius of the tire.

[0044] Technical effects and advantages of the present invention:

[0045] The present invention first uses a high-definition camera and a laser radar to collect visual data and depth information, and after graying, median filtering and frame difference processing, it identifies moving and stationary obstacles, and simplifies obstacles through an ellipsoid model. Secondly, based on the A* algorithm, the optimal path is generated, the environment is divided into a grid map, obstacles are marked and the path cost is calculated. Finally, the path is converted into a robot motion instruction through a PID control algorithm, and the linear velocity and angular velocity are calculated to drive the robot along the path; the present invention can flexibly respond to the appearance of obstacles in complex and dynamic environments through the above method, and the use of the A* algorithm can effectively reduce the travel length, reduce energy consumption, and improve work efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a schematic diagram of the method flow of the present invention. DETAILED DESCRIPTION

[0047] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] See also Figure 1 As shown, the present invention provides a robot motion control method based on artificial intelligence, and the method is:

[0049] Step A1: Collect visual data around the machine, depth data obtained by the lidar, and robot status data obtained by the sensor;

[0050] In step A1, a high-definition camera is used to collect visual data, and the visual data includes the front image data and continuous frames during the robot's driving process; a laser radar is used to collect the robot's current depth data, and the depth data includes the point cloud coordinates and distance data of the robot's current position.

[0051] Step A2: identifying obstacles and obstacle position coordinates in the environment by analyzing the visual data and the depth data;

[0052] In step A2, a three-dimensional coordinate system is constructed with the robot's starting position as the origin, and the collected image data is grayed out. The specific calculation method for converting a color image into a gray image is as follows:

[0053] A(D)=0.2989R+0.5870G+0.1140B, where A(D) represents the converted gray image, R represents the red intensity value, G represents the green intensity value, and B represents the blue intensity value;

[0054] The converted gray image is subjected to a median filter to remove noise;

[0055] The frame difference is calculated in the gray image after denoising. The specific method for calculating the frame difference is:

[0056] B(x, y)=|A(D t+1 (x,y))-A(D t (x, y))|, where B(x, y) represents the frame difference, (x, y) represents the pixel coordinates, and A(D t+1 (x, y)) represents the gray image of the obstacle position at time t+1, A(D t (x, y)) is represented as the gray image of the obstacle position at time t, D t+1 (x, y) represents the image frame collected at time t+1, D t(x, y) represents the image frame collected at time t;

[0057] A threshold is set to judge the frame difference and a binary image is generated. The specific calculation method for generating a binary image is:

[0058] Where C(x, y) represents the pixel value in the binary image, B(x, y) represents the frame difference, T represents the threshold, 1 represents the motion state, and 0 represents the static state;

[0059] Then the frame difference is judged by the threshold value, and the obstacles are classified into moving obstacles and stationary obstacles;

[0060] The threshold calculation method is as follows:

[0061] T=α+bβ, where T represents the threshold, α represents the average value of the frame difference, and β represents the standard deviation of the frame difference;

[0062] Among them, α represents the average value of all pixel differences, N represents the total number of pixels in the image, and B i (x, y) represents the frame difference of the i-th pixel, and (x, y) represents the pixel coordinates;

[0063] Among them, β represents the standard deviation of all pixel differences, α represents the average value of all pixel differences, N represents the total number of pixels in the image, and B i (x, y) represents the frame difference of the i-th pixel, and (x, y) represents the pixel coordinates;

[0064] The obstacle is wrapped by an n-dimensional ellipsoid to simplify the obstacle. The calculation method of the obstacle surface wrapped by the n-dimensional ellipsoid is as follows:

[0065] where Γ(s) is continuous and has continuous partial derivatives, n is the dimension, and s u Represented as the u-th value of the coordinate in n-dimensional space, Represented as the center of the ellipsoid, p u Expressed as the index of the ellipsoid, r u Expressed as the semi-axis length of the ellipsoid;

[0066] Depend on The four parameters determine an n-dimensional ellipsoid, parameter r u Determines the size of the ellipsoid on each axis, parameter p u Determines the degree of squareness of the ellipsoid, p u The larger the value, the closer the ellipsoid is to a cube. Determine the position of the ellipsoid;

[0067] The specific method for calculating the distance the robot reaches the obstacle is:

[0068] Where L is the distance from the robot to the obstacle, and c is the speed of light, which is 3×10 8 m / s, t 0 It is expressed as the time difference between the emission of the laser pulse and the reception of the reflected pulse;

[0069] The angle calculation method between the robot and the obstacle is as follows:

[0070] Among them, θ represents the angle between the robot and the obstacle, y 2 Represented as the value of the obstacle on the y-axis, y 1 Expressed as the value of the robot on the y-axis, x 2 Represented as the value of the obstacle on the x-axis, x 1 It is expressed as the value of the robot on the x-axis;

[0071] A vector is constructed with the robot's position as the starting point and the obstacle's position as the end point. The angle is determined by calculating the angle between the vector and the reference direction. The specific method for determining the angle is as follows:

[0072] The robot's position is F 1 (x 1 ,y 1 ), the position of the obstacle is F 2 (x 2 ,y 2 ), the reference direction is the x-axis, and the calculation method of the vector is as follows:

[0073] in, Represented as a vector, y 2 Represented as the value of the obstacle on the y-axis, y 1 Expressed as the value of the robot on the y-axis, x 2 Represented as the value of the obstacle on the x-axis, x 1 It is expressed as the value of the robot on the x-axis;

[0074] Then the vector The angle with the positive direction of the x-axis is the angle θ between the robot and the obstacle;

[0075] The position coordinates of the obstacle are determined based on the angle and the distance from the robot to the obstacle.

[0076] Step A3: Generate a robot motion path based on the identified obstacle position coordinates;

[0077] In step A3, the robot's working environment is divided into discrete grids, and the center point of each grid is regarded as a node; according to the position and shape of the obstacle, the inaccessible grid is marked on the map; then an open list and a closed list are created, the open list stores the nodes to be processed, and the closed list stores the processed nodes; the robot's starting position and target position are marked on the map as the starting point and end point of the A* algorithm respectively; the total cost value from the starting position to the target position is calculated, and the calculation method of the total cost value is specifically:

[0078] g(d)=h(d)+q(d), where g(d) represents the total cost from the initial position through point d to the target position, h(d) represents the actual cost from the starting position to node d, and q(d) represents the estimated cost of the optimal path from node d to the target position;

[0079] The calculation method of the estimated cost value of the optimal path from point d to the target location is as follows:

[0080] Where q(d) represents the estimated cost of the optimal path from point d to the target location, x e Represented as the value of the target position on the x-axis, x 1 Expressed as the value of the robot on the x-axis, y e Represented as the value of the target position on the y-axis, y 1 It is expressed as the value of the robot on the y-axis;

[0081] The actual cost value calculation method from the initial point to point d is as follows:

[0082] Where h(d) represents the actual cost from the starting position to node d, h 0 It is represented by the initial cost from the starting position to node d, k is the length of the current section, and v 0 Indicates the speed designed for the current road section;

[0083] Select the node with the smallest g(d) value from the open list as the current node, remove it from the open list and add it to the closed list; examine all adjacent nodes of the current node; for each adjacent node; if the node is in the closed list, ignore it; if the node is not in the open list, calculate its g(d) value and add it to the open list;

[0084] If the node is already in the open list, but the path to the node via the current node has a smaller h(d) value, then update the h(d) value and parent node information of the node, and recalculate its g(d) value;

[0085] Repeat the above steps until the target node is found; starting from the target node, trace back according to the parent node information until returning to the starting node; connect the nodes passed in the tracing process in sequence to form the optimal path from the starting node to the target node.

[0086] Step A4: receiving the motion path of the robot, and then controlling the robot to perform motion according to the motion path;

[0087] In step A4, the robot accepts the optimal path for driving, selects PID parameters, which include proportional gain, integral gain and differential gain; calculates the error according to the PID parameters, and then calculates the control signal according to the error, converts the control signal into the speed reached by the robot tire, and then outputs the calculated robot tire speed to the robot control system to drive the robot to move along the path;

[0088] The speed reached by the left tire is calculated as follows:

[0089] Among them, v l Expressed as the speed reached by the left wheel, v 1 Expressed as linear velocity, L 1 It is represented as the distance from the robot to the target position, φ is represented as the angular velocity, and H is represented as the radius of the tire;

[0090] The speed reached by the right tire is calculated as follows:

[0091] v R Expressed as the speed reached by the right wheel, v 1 Expressed as linear velocity, L 1 It is represented as the distance from the robot to the target position, φ is represented as the angular velocity, and H is represented as the radius of the tire;

[0092] The specific calculation method of line speed is:

[0093] v 1 =Q 1 m 1 +Q 2 m 2 +Q 3 m 3 , where v 1 Expressed as linear velocity, Q 1 Expressed as proportional gain, Q 2 Expressed as integral gain, Q 3 Expressed as differential gain, m 1 Expressed as proportional error, m 2 Expressed as the integral error, m 3 Expressed as differential error;

[0094] The calculation method of angular velocity is as follows:

[0095] Where φ is the angular velocity, Q 1 Expressed as proportional gain, Q 2 Expressed as integral gain, Q 3 It is represented as differential gain, Δθ is the angular error between the current position of the robot and the target position, Δt is the sampling time interval, and Δθ 1 It is expressed as the angle deviation at the last sampling moment;

[0096] The calculation method of proportional error is as follows:

[0097] Among them, m 1 It is expressed as proportional error, Δx is the position error on the x-axis, and Δy is the position error on the y-axis;

[0098] Δx=x e -x 1 , where Δx represents the position error on the x-axis, x e Represented as the value of the target position on the x-axis, x 1 It is expressed as the value of the robot on the x-axis;

[0099] Δy=y e -y 1 , where Δy represents the position error on the y-axis, y e Represented as the value of the target position on the y-axis, y 1 It is expressed as the value of the robot on the y-axis;

[0100] The calculation method of the integral error is as follows:

[0101] m 2 =m 20 +m 1 Δt, where m 2 Expressed as the integral error, m 20 Expressed as the integrated error of the previous moment, m 1 It is expressed as proportional error, and Δt is the sampling time interval;

[0102] The calculation method of differential error is as follows:

[0103] m 3 It is expressed as differential error, Δt is the sampling time interval, m 1 Expressed as proportional error, m 10 It is expressed as the last proportional error;

[0104] During the execution process, the robot status is updated in real time and the new error is calculated; if the error exceeds the set threshold, adjustments are performed: the actual position of the robot is monitored and compared with the target position, and the PID output is recalculated.

[0105] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A robot motion control method based on artificial intelligence, characterized in that: include: Step A1: Collect visual data around the machine, depth data obtained by the lidar, and robot status data obtained by the sensor; Step A2: identifying obstacles and obstacle position coordinates in the environment by analyzing the visual data and the depth data; Step A3: Generate a robot motion path based on the identified obstacle position coordinates; Step A4: receiving the motion path of the robot, and then controlling the robot to execute motion according to the motion path.

2. The robot motion control method based on artificial intelligence according to claim 1, characterized in that: In step A1, a high-definition camera is used to collect visual data, and the visual data includes the front image data and continuous frames during the robot's driving process; a laser radar is used to collect the robot's current depth data, and the depth data includes the point cloud coordinates and distance data of the robot's current position.

3. The robot motion control method based on artificial intelligence according to claim 1, characterized in that: In the step A2, a three-dimensional coordinate system is constructed with the robot's starting position as the origin, and the collected image data is grayed out; The converted gray image is subjected to a median filter to remove noise; The frame difference is calculated in the denoised gray image. The specific method for calculating the frame difference is: B(x, y)=|A(D t+1 (x,y))-A(D t (x, y))|, where B(x, y) represents the frame difference, (x, y) represents the pixel coordinates, and A(D t+1 (x, y)) represents the gray image of the obstacle position at time t+1, A(D t (x, y)) is represented as the gray image of the obstacle position at time t, D t+1 (x, y) represents the image frame collected at time t+1, D t (x, y) represents the image frame collected at time t; A threshold is set to judge the frame difference and a binary image is generated. The specific calculation method for generating a binary image is: Where C(x, y) represents the pixel value in the binary image, B(x, y) represents the frame difference, T represents the threshold, 1 represents the motion state, and 0 represents the static state; The frame difference is judged by the threshold and the obstacles are classified into moving obstacles and stationary obstacles.

4. The robot motion control method based on artificial intelligence according to claim 3 is characterized in that: The obstacle is wrapped by an n-dimensional ellipsoid to simplify the obstacle. The calculation method of the obstacle surface wrapped by the n-dimensional ellipsoid is as follows: where Γ(s) is continuous and has continuous partial derivatives, n is the dimension, and s u Represented as the u-th value of the coordinate in n-dimensional space, Represented as the center of the ellipsoid, p u Expressed as the index of the ellipsoid, r u Expressed as the semi-axis length of the ellipsoid; Depend on The four parameters determine an n-dimensional ellipsoid, parameter r u Determines the size of the ellipsoid on each axis, parameter p u Determines the degree of squareness of the ellipsoid, p u The larger the value, the closer the ellipsoid is to a cube. Determines the position of the ellipsoid.

5. The robot motion control method based on artificial intelligence according to claim 3 is characterized in that: A vector is constructed with the robot's position as the starting point and the obstacle's position as the end point. The angle is determined by calculating the angle between the vector and the reference direction. The specific method for determining the angle is as follows: The position of the robot is F1 (x1, y1), the position of the obstacle is F2 (x2, y2), the reference direction is the x-axis, and the vector calculation method is as follows: in, Represented as a vector, y2 represents the value of the obstacle on the y-axis, y1 represents the value of the robot on the y-axis, x2 represents the value of the obstacle on the x-axis, and x1 represents the value of the robot on the x-axis; Then the vector The angle with the positive direction of the x-axis is the angle θ between the robot and the obstacle; The position coordinates of the obstacle are determined based on the angle and the distance from the robot to the obstacle.

6. The robot motion control method based on artificial intelligence according to claim 1, characterized in that: In step A3, the robot's working environment is divided into discrete grids, and the center point of each grid is regarded as a node; the inaccessible grids are marked on the map according to the position and shape of the obstacles; and an open list and a closed list are created, the open list stores the nodes to be processed, and the closed list stores the processed nodes; Mark the robot's starting position and target position on the map as the starting point and end point of the A* algorithm respectively; calculate the total cost from the starting position to the target position. The calculation method of the total cost value is as follows: g(d)=h(d)+q(d), where g(d) represents the total cost from the initial position through point d to the target position, h(d) represents the actual cost from the starting position to node d, and q(d) represents the estimated cost of the optimal path from node d to the target position; Select the node with the smallest g(d) value from the open list as the current node, remove it from the open list and add it to the closed list; examine all adjacent nodes of the current node; for each adjacent node; if the node is in the closed list, ignore it; if the node is not in the open list, calculate its g(d) value and add it to the open list; If the node is already in the open list, but the path to the node via the current node has a smaller h(d) value, then update the h(d) value and parent node information of the node, and recalculate its g(d) value; Repeat the above steps until the target node is found; starting from the target node, trace back according to the parent node information until returning to the starting node; connect the nodes passed in the tracing process in sequence to form the optimal path from the starting node to the target node.

7. The robot motion control method based on artificial intelligence according to claim 1, characterized in that: In step A4, the robot accepts the optimal path for driving, selects PID parameters, which include proportional gain, integral gain and differential gain; calculates the error according to the PID parameters, and then calculates the control signal according to the error, converts the control signal into the speed reached by the robot tire, and then outputs the calculated robot tire speed to the robot control system to drive the robot to move along the path; The speed reached by the left tire is calculated as follows: Among them, v l represents the speed reached by the left wheel, v1 represents the linear velocity, L1 represents the distance from the robot to the target position, φ represents the angular velocity, and H represents the radius of the tire; The speed reached by the right tire is calculated as follows: v R It represents the speed reached by the right wheel, v1 represents the linear speed, L1 represents the distance from the robot to the target position, φ represents the angular velocity, and H represents the radius of the tire.

8. The robot motion control method based on artificial intelligence according to claim 7 is characterized in that: The specific method for calculating the linear velocity is: v1=Q1m1+Q2m2+Q3m3, where v1 represents the linear velocity, Q1 represents the proportional gain, Q2 represents the integral gain, Q3 represents the differential gain, m1 represents the proportional error, m2 represents the integral error, and m3 represents the differential error; The calculation method of angular velocity is as follows: Among them, φ represents the angular velocity, Q1 represents the proportional gain, Q2 represents the integral gain, Q3 represents the differential gain, Δθ represents the angular error between the current position of the robot and the target position, Δt represents the sampling time interval, and Δθ1 represents the angular deviation at the last sampling moment; During the execution process, the robot status is updated in real time and the new error is calculated; if the error exceeds the set threshold, adjustments are performed: the actual position of the robot is monitored and compared with the target position, and the PID output is recalculated.

Images