An improved method for inspection unmanned vehicle based on dynamic path planning

Abstract

An improved method for inspection unmanned vehicle based on dynamic path planning, through laser radar, odometer and IMU information fusion, the position of the inspection unmanned vehicle is obtained and the speed of the inspection unmanned vehicle is sampled; the speed of the moving obstacle is collected, whether the inspection unmanned vehicle and the moving obstacle will have the possibility of driving in the same direction is judged by judging the included angle between the moving speed direction of the inspection unmanned vehicle and the moving speed direction of the moving obstacle, and the advancing direction of the inspection unmanned vehicle is optimized; then the trajectory space of the inspection unmanned vehicle is generated and the trajectory space is evaluated, whether the inspection unmanned vehicle and the obstacle will collide is judged, if yes, the trajectory route is regenerated; if not, the optimal trajectory is selected and the optimal path is generated, and the path regularization is ended. The real-time dynamic obstacle avoidance of the inspection unmanned vehicle to the moving obstacle and the selection of the optimal trajectory are realized, the algorithm complexity is low, the realization is simple, and the method can be used in the inspection work of the inspection unmanned vehicle in the actual industrial field.

Description

Technical Field

[0001] This invention relates to the field of obstacle avoidance technology for unmanned vehicles, and specifically to an improved method for inspection unmanned vehicles based on dynamic path planning. Background Technology

[0002] In recent years, path planning and navigation positioning technologies for unmanned inspection vehicles have gradually matured. However, in the process of obstacle avoidance path planning for unmanned vehicles, achieving real-time obstacle avoidance and path planning for dynamic obstacles has become a technical challenge, compared to obstacle avoidance for fixed obstacles. The DWA algorithm, with its advantages of low computational complexity and real-time obstacle avoidance, has begun to be widely used in the development and application of local path planning. The algorithm's implementation principle involves introducing an evaluation function to evaluate and score the generated trajectory, thereby selecting the optimal trajectory. However, compared to global planning algorithms such as A* and Dijkstra, this algorithm only selects the best path at the current moment when avoiding dynamic obstacles in real time, rather than the best path based on global considerations. This results in the final generated trajectory not being the globally optimal trajectory.

[0003] Although the algorithm normalizes the evaluation function to avoid the situation where the final selected route trajectory is not the optimal trajectory because a certain item is not dominant in the evaluation function, it only processes and reduces the error based on the trajectory route obtained by the DWA algorithm, without making any substantial changes. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention proposes an improved method for inspection unmanned vehicles based on dynamic path planning. The evaluation function is improved by adding a barr(v) evaluation function as the resultant velocity to the evaluation function, specifically addressing the prediction of dynamic obstacles during the path planning process of the inspection unmanned vehicle. c ,v b This allows for the optimization of errors in dynamic obstacle avoidance and path planning, resulting in a more globally-oriented optimal trajectory and effectively solving the aforementioned technical problems.

[0005] This invention is achieved through the following technical solution:

[0006] An improved method for inspection unmanned vehicles based on dynamic path planning is proposed. This method fuses information from LiDAR, odometer, and IMU to obtain the position of the inspection vehicle and samples its speed. It also collects the speed of moving obstacles and determines the possibility of the vehicle and obstacle traveling in the same direction by judging the angle between their respective speed directions. The method optimizes the vehicle's forward direction, generates its trajectory space, and evaluates it to determine if a collision is likely. If a collision occurs, a new trajectory is generated; otherwise, the optimal trajectory is selected and the optimal path is generated. The specific operation is as follows:

[0007] Step 1: Collect data and fuse the data;

[0008] The system fuses information from lidar, odometer, and IMU mounted on the vehicle, analyzes and determines the location of the unmanned inspection vehicle based on the collected data, and samples the speed of the unmanned inspection vehicle.

[0009] Step 2: Determine the direction of travel of the unmanned inspection vehicle and the moving obstacles;

[0010] By collecting the speed of moving obstacles using lidar, and by measuring the angle between the direction of the unmanned vehicle's speed and the direction of the moving obstacle's speed, it can be determined whether the unmanned vehicle and the moving obstacle are likely to travel in the same direction.

[0011] Step 3: Optimize the forward direction of the inspection drone based on the judgment results;

[0012] When the angle between the direction of the resultant velocity of the inspection vehicle and the moving obstacle and the direction of the moving obstacle's velocity is an acute angle, it is determined that the inspection vehicle and the moving obstacle may travel in the same direction; the forward direction of the inspection vehicle needs to be optimized so that the angle between the direction of the resultant velocity of the inspection vehicle and the moving obstacle and the direction of the moving obstacle's velocity is an obtuse angle.

[0013] Step 4: Generate the trajectory space of the inspection unmanned vehicle and evaluate the trajectory space;

[0014] Determine whether the unmanned inspection vehicle will collide with an obstacle. If a collision is expected, regenerate the trajectory route; if no collision is expected, select the optimal trajectory and generate the optimal path.

[0015] Step 5: When the inspection drone reaches the target point, select the optimal trajectory and generate the optimal path, and end the path planning.

[0016] Furthermore, the sampling of the speed of the inspection drone described in step 1 is specifically performed as follows: Given that the radius of the inspection drone's wheels is r, and the speed of the inspection drone is the average speed of its wheels, there is a rotational speed measuring unit on each of the left and right wheels in the chassis of the inspection drone. The rotational speed of the two wheels can be obtained through these units, thus yielding the inspection drone's travel speed. The specific formula is as follows:

[0017] c = 2πr

[0018] v l =a l ×c

[0019] v r =a r ×c

[0020]

[0021] In the above formula, c is the circumference of the wheel, which is the distance traveled by the wheel in one revolution; al and ar are the rotational speeds of the left and right wheels; vl and vr are the speeds of the left and right wheels; and vc is the speed of the unmanned inspection vehicle at this moment.

[0022] Furthermore, the specific operation method for collecting the speed of moving obstacles based on the LiDAR in step 2 is as follows: the inspection unmanned vehicle scans and establishes a costmap grid map using LiDAR. The LiDAR probe continuously emits lasers in all directions. The time interval from the emission of the laser to its contact with the moving obstacle and its return to the receiver is t. The position of the moving obstacle scanned by the laser on the grid map is set as (x, y). By continuously emitting lasers in all directions and collecting data, the positions reached by a finite number of moving obstacles after traveling can be obtained. The more times the position is sampled, the more accurate the collected speed will be. The speed collection formula is as follows:

[0023]

[0024]

[0025]

[0026] In the formula, Xti is the distance between the position at the previous time and the position at the next time, that is, the distance traveled by the moving obstacle in time ti, vti is the instantaneous speed of the moving obstacle at time ti, and vb is the average speed of the moving obstacle. The more positions of the moving obstacle are collected, the more accurate the value of vb will be, and the smaller the error will be.

[0027] This allows us to determine the speed and direction of the unmanned inspection vehicle and the moving obstacle. Since the speed is a vector, the extensions of the two speed directions will form an angle when they are translated vertically in the same two-dimensional plane. This angle can be used to determine whether the unmanned inspection vehicle and the moving obstacle are likely to travel in the same direction.

[0028] Furthermore, step 3, which involves optimizing the forward direction of the inspection drone based on the judgment result, is specifically implemented as follows:

[0029] Step 1: Based on the known magnitude and direction of the speed of the unmanned inspection vehicle and the moving obstacle, establish a rectangular coordinate system with the intersection point as the origin at the angle, and use the parallelogram rule to calculate the magnitude and direction of the resultant speed of the two.

[0030] Step 2: Determine if they are traveling in the same direction: In the rectangular coordinate system established at the angle, if the directions of the velocities of the two objects are located in adjacent or the same quadrant in the four quadrants of the coordinate system (I, II, III, IV), then it is considered that there is a possibility of traveling in the same direction. In this case, the angle between the resultant velocity of the two objects and the velocity of the moving obstacle must be an acute angle.

[0031] Step 3: When there is a possibility of traveling in the same direction, in order to avoid the risk of collision caused by traveling in the same direction, given the magnitude and direction of the speed of the moving obstacle, the direction of the inspection unmanned vehicle is optimized and changed. According to the parallelogram law, the direction of the resultant velocity is changed, thereby changing the angle between the resultant velocity and the moving obstacle. By optimizing and controlling the direction of the inspection unmanned vehicle, this angle is adjusted to an obtuse angle of less than 180°, so that the inspection unmanned vehicle and the moving obstacle will not travel in the same direction, thereby achieving the purpose of avoiding collision.

[0032] Furthermore, the optimization of the angle of the inspection unmanned vehicle is specifically implemented as follows: For the prediction of moving obstacles during the path planning process of the inspection unmanned vehicle, the prediction algorithm adds a resultant velocity evaluation function barr(vc,vb) to the evaluation function of the inspection unmanned vehicle. That is, it evaluates the angle between the resultant velocity of the inspection unmanned vehicle and the moving obstacle, and the moving obstacle's velocity. The evaluation function of the inspection unmanned vehicle is:

[0033] G(v,w)=[σ·heading(v,w)+β·dist(v,w)+γ·velocity(v,w)]·barr(v c ,v b )

[0034] Normalizing the evaluation function yields:

[0035]

[0036]

[0037]

[0038]

[0039] In the formula, barr(v c ,v b ) is the evaluation function for the resultant velocity, which evaluates the angle between the resultant velocity of the inspection unmanned vehicle and the velocity of the moving obstacle and the velocity of the moving obstacle.

[0040] Furthermore, by continuously scanning the environment and recording the moving speed of the moving obstacle using the aforementioned lidar, when the angle between the combined speed of the inspection drone and the moving obstacle's speed and the moving obstacle's speed is an acute angle, the inspection drone's travel speed v is taken as... c1 The speed v of the moving obstacle b1 The resultant velocity is v h1 Take v h1 The normal vector is v h2 The improved inspection unmanned vehicle's travel speed v c2 With -v b1 The resultant velocity is v h2 Find the improved speed v of the unmanned inspection vehicle. c2 The rotational speeds of the four wheels of the inspection vehicle can be obtained from the target speed, and the direction of travel of the inspection vehicle can be optimized by controlling the rotational speeds of the four wheels using a Raspberry Pi. The optimization formula is as follows:

[0041]

[0042]

[0043]

[0044]

[0045] In the formula, v c1 To inspect the speed of the unmanned vehicle, v b1 v is the speed at which the dynamic obstacle moves. h1 For v c1 With v b1 The resultant velocity, v g2 For v h1 The normal vector, v c2 To improve the travel speed of the unmanned inspection vehicle.

[0046] Furthermore, the specific operation method for generating the trajectory space of the inspection unmanned vehicle in step 4 is as follows: according to the optimized control of the travel direction of the inspection unmanned vehicle in step 3, the safety hazard of collision when traveling in the same direction as the moving obstacle is avoided. While avoiding the moving obstacle, a new travel route is planned. When the inspection unmanned vehicle is traveling according to the planned new travel route, it can continue to scan the moving obstacle to plan a travel route to avoid the obstacle. The new travel route generated by the obstacle avoidance of the inspection unmanned vehicle is the trajectory space.

[0047] Furthermore, the evaluation of the trajectory route in the generated trajectory space in step 4 is specifically carried out as follows: it is determined whether the driving route of the inspection unmanned vehicle generated in the trajectory space will collide with moving obstacles. If a collision occurs, step 3 is executed again to generate the latest trajectory space; if no collision occurs, the trajectory route with the shortest route and the best obstacle avoidance effect is selected.

[0048] Furthermore, the specific operation method for generating the optimal path in step 5 is as follows: based on the optimal trajectory selected in step 4, evaluate and judge whether the optimal trajectory of the inspection unmanned vehicle has reached the target point. If it has reached the target point, generate the optimal path; otherwise, generate the trajectory space map and repeat steps 3 and 4 until the inspection unmanned vehicle safely reaches the target point.

[0049] Beneficial effects

[0050] The present invention proposes an improved method for inspection unmanned vehicles based on dynamic path planning, which has the following advantages compared with existing technologies:

[0051] (1) This technical solution optimizes and improves the evaluation function. Specifically, during the path planning process of the unmanned inspection vehicle, it compensates for the error caused by the algorithm's evaluation function when evaluating the generated trajectory. An evaluation function barr(v) is added to the evaluation function of the unmanned inspection vehicle to represent the resultant velocity. c b) This involves evaluating the angle between the direction of the resultant velocity of the inspection drone and the direction of the moving obstacle's velocity and the direction of the moving obstacle's velocity. This reduces the error caused by the algorithm only selecting the best path at the current moment instead of the best path based on global considerations when avoiding dynamic obstacles in real time. The resulting trajectory is not the globally optimal trajectory. This invention improves the original algorithm, making up for the shortcomings of the original algorithm in trajectory selection. It introduces a resultant velocity evaluation function into the original evaluation function and adopts a new constraint method for trajectory selection. This realizes the optimization of real-time dynamic obstacle avoidance and optimal trajectory selection for inspection drones. Moreover, the algorithm has low computational complexity, is simple to implement, and can be applied to the inspection work of inspection drones in the actual industrial field.

[0052] (2) This technical solution uses the data from the odometer, IMU and lidar sensor to perform positioning, which makes the system more stable and reduces the measurement error of the odometer information due to the error of the actuator and possible sideslip in actual application. Attached Figure Description

[0053] Figure 1 This is a detailed flowchart of the present invention.

[0054] Figure 2 This is a schematic diagram of the optimized forward direction of the unmanned inspection vehicle in this invention.

[0055] Figure 3 This is a schematic diagram illustrating the generation of the optimal path in this invention. Detailed Implementation

[0056] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0057] Example 1:

[0058] like Figure 1 As shown, an improved method for inspection unmanned vehicles based on dynamic path planning is proposed. This method fuses information from LiDAR, odometer, and IMU to obtain the position of the inspection unmanned vehicle and samples its speed. It also collects the speed of moving obstacles and determines the possibility of the vehicle and obstacle traveling in the same direction by judging the angle between their respective speed directions. The method optimizes the vehicle's forward direction, generates its trajectory space, and evaluates it to determine if a collision is likely. If a collision occurs, a new trajectory is generated; otherwise, the optimal trajectory is selected and the optimal path is generated. The specific operation is as follows:

[0059] Step 1: Collect data and fuse the data;

[0060] The system integrates information from lidar, odometer, and IMU mounted on the vehicle, analyzes and determines the location of the unmanned inspection vehicle based on the collected data, and samples the speed of the unmanned inspection vehicle.

[0061] The sampling of the speed of the inspection drone is performed as follows: Given that the radius of the inspection drone's wheels is r, and the speed of the inspection drone is the average speed of its wheels, there is a rotational speed measuring unit on each of the left and right wheels in the chassis of the inspection drone. The rotational speed of the two wheels can be obtained through these units, thus determining the vehicle's speed. The specific formula is as follows:

[0062] c = 2πr

[0063] v l =a l ×c

[0064] v r =a r ×c

[0065]

[0066] In the above formula, c is the circumference of the wheel, which is the distance traveled by the wheel in one revolution; al and ar are the rotational speeds of the left and right wheels; vl and vr are the speeds of the left and right wheels; and vc is the speed of the unmanned inspection vehicle at this moment.

[0067] Step 2: Determine the direction of travel of the unmanned inspection vehicle and the moving obstacles;

[0068] By collecting the speed of moving obstacles using lidar, and by measuring the angle between the direction of the unmanned vehicle's speed and the direction of the moving obstacle's speed, it can be determined whether the unmanned vehicle and the moving obstacle are likely to travel in the same direction.

[0069] The specific operation method for collecting the speed of moving obstacles based on lidar is as follows: The inspection unmanned vehicle scans and establishes a costmap grid map using lidar. The lidar probe continuously emits lasers in all directions. The time interval from the emission of the laser to its contact with the moving obstacle and its return to the receiver is t. The position of the moving obstacle scanned by the laser on the grid map is set as (x, y). By continuously emitting lasers in all directions and collecting data, the positions reached by a finite number of moving obstacles after traveling can be obtained. The more times the position is sampled, the more accurate the collected speed will be. The speed collection formula is as follows:

[0070]

[0071]

[0072]

[0073] In the formula, Xti is the distance between the position at the previous time and the position at the next time, that is, the distance traveled by the moving obstacle in time ti, vti is the instantaneous speed of the moving obstacle at time ti, and vb is the average speed of the moving obstacle. The more positions of the moving obstacle are collected, the more accurate the value of vb will be, and the smaller the error will be.

[0074] This allows us to determine the speed and direction of the unmanned inspection vehicle and the moving obstacle. Since the speed is a vector, the extensions of the two speed directions will form an angle when they are translated vertically in the same two-dimensional plane. This angle can be used to determine whether the unmanned inspection vehicle and the moving obstacle are likely to travel in the same direction.

[0075] Step 3: Optimize the forward direction of the inspection drone based on the judgment results;

[0076] When the angle between the direction of the combined velocity of the inspection vehicle and the moving obstacle and the direction of the moving obstacle's velocity is acute, it is determined that the inspection vehicle and the moving obstacle may travel in the same direction; the forward direction of the inspection vehicle needs to be optimized so that the angle between the direction of the combined velocity of the inspection vehicle and the moving obstacle and the direction of the moving obstacle's velocity is obtuse.

[0077] like Figure 2 As shown, the forward direction of the inspection unmanned vehicle is optimized based on the judgment results. The specific operation method is as follows:

[0078] Step 1: Based on the known magnitude and direction of the speed of the unmanned inspection vehicle and the moving obstacle, establish a rectangular coordinate system with the intersection point as the origin at the angle, and use the parallelogram rule to calculate the magnitude and direction of the resultant speed of the two.

[0079] Step 2: Determine if they are traveling in the same direction: In the rectangular coordinate system established at the angle, if the directions of the velocities of the two objects are located in adjacent or the same quadrant in the four quadrants of the coordinate system (I, II, III, IV), then it is considered that there is a possibility of traveling in the same direction. In this case, the angle between the resultant velocity of the two objects and the velocity of the moving obstacle must be an acute angle.

[0080] Step 3: When there is a possibility of traveling in the same direction, in order to avoid the risk of collision caused by traveling in the same direction, given the magnitude and direction of the speed of the moving obstacle, the direction of the inspection unmanned vehicle is optimized and changed. According to the parallelogram law, the direction of the resultant velocity is changed, thereby changing the angle between the resultant velocity and the moving obstacle. By optimizing and controlling the direction of the inspection unmanned vehicle, this angle is adjusted to an obtuse angle of less than 180°, so that the inspection unmanned vehicle and the moving obstacle will not travel in the same direction, thereby achieving the purpose of avoiding collision.

[0081] To optimize the angle of the inspection drone, the specific operation method is as follows: For the prediction of moving obstacles during the path planning process of the inspection drone, the prediction algorithm adds a resultant velocity evaluation function barr(vc,vb) to the evaluation function of the inspection drone. This function evaluates the angle between the resultant velocity of the inspection drone and the moving obstacle, and the moving obstacle's velocity. The evaluation function of the inspection drone is as follows:

[0082] G(v,w)=[σ·heading(v,w)+β·dist(v,w)+γ·velocity(v,w)]·barr(v c ,v b )

[0083] Normalizing the evaluation function yields:

[0084]

[0085]

[0086]

[0087]

[0088] In the formula, barr(v c ,v b ) is the evaluation function for the resultant velocity, which evaluates the angle between the resultant velocity of the inspection unmanned vehicle and the velocity of the moving obstacle and the velocity of the moving obstacle.

[0089] The moving speed of the inspection drone is recorded by continuously scanning the environment with lidar. When the angle between the combined speed of the inspection drone and the moving obstacle and the moving obstacle's speed is acute, the inspection drone's speed v is taken. c1 The speed v of the moving obstacle b1 The resultant velocity is v h1 Take v h1 The normal vector is v h2 The improved inspection unmanned vehicle's travel speed v c2 With -v b1 The resultant velocity is v h2 Find the improved speed v of the unmanned inspection vehicle. c2 The rotational speeds of the four wheels of the inspection vehicle can be obtained from the target speed, and the direction of travel of the inspection vehicle can be optimized by controlling the rotational speeds of the four wheels using a Raspberry Pi. The optimization formula is as follows:

[0090]

[0091]

[0092]

[0093]

[0094] In the formula, v c1 To inspect the speed of the unmanned vehicle, v b1 v is the speed at which the dynamic obstacle moves. h1 For v c1 With v b1 The resultant velocity, v h2 For v h1 The normal vector, v c2 To improve the travel speed of the unmanned inspection vehicle.

[0095] Step 4: Generate the trajectory space of the inspection unmanned vehicle and evaluate the trajectory space;

[0096] The system determines whether the unmanned inspection vehicle will collide with an obstacle. If a collision is expected, a new trajectory is generated. If no collision is expected, the optimal trajectory is selected and the optimal path is generated. The specific operation is as follows: the unmanned inspection vehicle's direction of travel is optimized according to step 3 to avoid the safety hazard of collision when traveling in the same direction as a moving obstacle. While avoiding the moving obstacle, a new driving route is planned. When the unmanned inspection vehicle is traveling along the planned new driving route, it can continue to scan for moving obstacles to plan a driving route to avoid obstacles. The new driving route generated by the unmanned inspection vehicle's obstacle avoidance is the trajectory space.

[0097] The generated trajectory in the trajectory space is evaluated. Specifically, it is determined whether the driving route of the inspection unmanned vehicle generated in the trajectory space will collide with moving obstacles. If a collision occurs, step 3 is executed again to generate the latest trajectory space. If no collision occurs, the trajectory with the shortest route and the best obstacle avoidance effect is selected.

[0098] Step 5: When the inspection vehicle reaches the target point, select the optimal trajectory and generate the optimal path. Specifically, evaluate the optimal trajectory selected in Step 4 to determine if the optimal trajectory of the inspection vehicle has reached the target point. If it has, generate the optimal path as shown in the image. Figure 3 The optimal path is shown; if the target is not reached, a trajectory space map is generated, and steps 3 and 4 are repeated until the inspection vehicle safely reaches the target point; the path planning ends.

[0099] Matters not covered in this invention are common knowledge.

[0100] The above embodiments are only for illustrating the technical concept and features of the present invention, and are intended to enable those skilled in the art to understand the content of the present invention and implement it accordingly. They should not be construed as limiting the scope of protection of the present invention. All equivalent changes or modifications made in accordance with the spirit and essence of the present invention should be covered within the scope of protection of the present invention.

Claims

1. An improved method for inspection unmanned vehicles based on dynamic path planning, characterized in that: The location of the inspection vehicle is obtained by fusing information from lidar, odometer, and IMU, and the speed of the inspection vehicle is sampled. The speed of moving obstacles is collected, and the angle between the direction of the inspection vehicle's (VV) speed and the direction of the obstacle's speed is used to determine whether the VV and the obstacle are likely to travel in the same direction. The forward direction of the VV is optimized, and then the trajectory space of the VV is generated and evaluated to determine whether a collision will occur between the VV and the obstacle. If a collision occurs, the trajectory route is regenerated; if not, the optimal trajectory is selected and the optimal path is generated. The specific operation method is as follows: Step 1: Collect data and fuse the data; The system fuses information from lidar, odometer, and IMU mounted on the vehicle, analyzes and determines the location of the unmanned inspection vehicle based on the collected data, and samples the speed of the unmanned inspection vehicle. Step 2: Determine the direction of travel of the unmanned inspection vehicle and the moving obstacles; By collecting the speed of moving obstacles using lidar, and by measuring the angle between the direction of the unmanned vehicle's speed and the direction of the moving obstacle's speed, it can be determined whether the unmanned vehicle and the moving obstacle are likely to travel in the same direction. Step 3: Optimize the forward direction of the inspection drone based on the judgment results; When the angle between the direction of the resultant velocity of the inspection vehicle and the moving obstacle and the direction of the moving obstacle's velocity is an acute angle, it is determined that the inspection vehicle and the moving obstacle may travel in the same direction; the forward direction of the inspection vehicle needs to be optimized so that the angle between the direction of the resultant velocity of the inspection vehicle and the moving obstacle and the direction of the moving obstacle's velocity is an obtuse angle. Changing the travel direction of the inspection drone alters the angle between the resultant velocity and the moving obstacle. This includes: predicting moving obstacles during the path planning process of the inspection drone. The prediction algorithm adds a resultant velocity evaluation function `barr(vc, vb)` to the evaluation function of the inspection drone. Specifically, it evaluates the angle between the resultant velocity of the inspection drone and the moving obstacle, and the moving obstacle's velocity. The evaluation function of the inspection drone is as follows: ； Normalizing the evaluation function yields: ； ； ； ； In the formula, The evaluation function is the resultant velocity, which evaluates the angle between the resultant velocity of the inspection unmanned vehicle and the velocity of the moving obstacle and the velocity of the moving obstacle. Step 4: Generate the trajectory space of the inspection unmanned vehicle and evaluate the trajectory space; Determine whether the unmanned inspection vehicle will collide with an obstacle. If a collision is expected, regenerate the trajectory route; if no collision is expected, select the optimal trajectory and generate the optimal path. Step 5: When the inspection drone reaches the target point, select the optimal trajectory and generate the optimal path, and end the path planning.

2. The improved method for inspection unmanned vehicles based on dynamic path planning according to claim 1, characterized in that: Step 1 involves sampling the speed of the unmanned inspection vehicle. The specific operation is as follows: Given that the radius of the unmanned inspection vehicle's wheels is r, and the vehicle's speed is the average speed of its wheels, the chassis of the unmanned inspection vehicle has a rotational speed measurement unit on each of the left and right wheels. These units can be used to obtain the rotational speed of the two wheels at that moment, thus yielding the vehicle's travel speed. The specific formula is as follows: ； ； ； In the above formula, c is the circumference of the wheel, that is, the distance the wheel travels in one revolution, and a l a r v represents the rotational speed of the left and right wheels. l v r v represents the speed of the left and right wheels. c The speed of the unmanned vehicle being inspected at this moment.

3. The improved method for inspection unmanned vehicles based on dynamic path planning according to claim 1, characterized in that: Step 2, which involves collecting the speed of moving obstacles using LiDAR, is performed as follows: The unmanned inspection vehicle scans and creates a costmap grid map using LiDAR. The LiDAR probe continuously emits lasers in all directions. The time interval from the emission of the laser to its contact with the moving obstacle and its return to the receiver is t. The position of the moving obstacle scanned by the laser on the grid map is set as (x, y). By continuously emitting lasers in all directions and collecting data, the positions reached by a finite number of moving obstacles after they have traveled can be obtained. The more times the position is sampled, the more accurate the collected speed will be. The speed collection formula is as follows: ； ； ； In the formula, X ti v represents the distance between the previous and next time position, i.e., the distance traveled by moving the obstacle within time ti. ti Let v be the instantaneous velocity of the moving obstacle at time ti. b v represents the average speed of the moving obstacle. The more locations of the moving obstacles are collected, the higher v becomes. b The more accurate the value, the smaller the error. This allows us to determine the speed and direction of the unmanned inspection vehicle and the moving obstacle. Since the speed is a vector, the extensions of the two speed directions will form an angle when they are translated vertically in the same two-dimensional plane. Based on this angle, we can determine whether the unmanned inspection vehicle and the moving obstacle are likely to travel in the same direction.

4. An improved method for inspection unmanned vehicles based on dynamic path planning according to claim 1, characterized in that: Step 3, which involves optimizing the forward direction of the unmanned inspection vehicle based on the judgment result, is specifically implemented as follows: Step 1: Based on the known magnitude and direction of the speed of the unmanned inspection vehicle and the moving obstacle, establish a rectangular coordinate system with the intersection point as the origin at the angle, and use the parallelogram rule to calculate the magnitude and direction of the resultant speed of the two. Step 2: Determine if they are traveling in the same direction: In the rectangular coordinate system established at the angle, if the directions of the velocities of the two objects are located in adjacent or the same quadrant in the four quadrants of the coordinate system (I, II, III, IV), then it is considered that there is a possibility of traveling in the same direction. In this case, the angle between the resultant velocity of the two objects and the velocity of the moving obstacle must be an acute angle. Step 3: When there is a possibility of traveling in the same direction, in order to avoid the risk of collision caused by traveling in the same direction, given the magnitude and direction of the speed of the moving obstacle, the direction of the inspection unmanned vehicle is changed by optimizing and changing the direction of the resultant velocity according to the parallelogram law, thereby changing the angle between the resultant velocity and the moving obstacle. By optimizing and controlling the direction of the inspection unmanned vehicle, this angle is adjusted to an obtuse angle, so that the inspection unmanned vehicle and the moving obstacle will not travel in the same direction, thereby achieving the purpose of avoiding collision.

5. An improved method for inspection unmanned vehicles based on dynamic path planning according to claim 1, characterized in that: The laser radar continuously scans the environment and records the moving speed of moving obstacles. When the angle between the combined speed of the inspection drone and the moving obstacle's speed and the moving obstacle's speed is acute, the inspection drone's speed is taken. Speed ​​of travel with dynamic obstacles The resultant velocity is ,Pick The normal vector is This improved the travel speed of the unmanned inspection vehicle. and The resultant velocity is Find the improved speed of the unmanned inspection vehicle. The rotational speeds of the four wheels of the inspection vehicle can be obtained from the target speed, and the direction of travel of the inspection vehicle can be optimized by controlling the rotational speeds of the four wheels using a Raspberry Pi. The optimization formula is as follows: ； ； ； ； In the formula, To inspect the speed of the unmanned vehicle, The speed at which the obstacle moves. for and The combined velocity, for The normal vector, To improve the travel speed of the unmanned inspection vehicle.

6. An improved method for inspection unmanned vehicles based on dynamic path planning according to claim 5, characterized in that: The specific operation method for generating the trajectory space of the inspection unmanned vehicle in step 4 is as follows: Based on the optimization of the travel direction of the inspection unmanned vehicle in step 3, the safety hazards of collision when traveling in the same direction as moving obstacles are avoided. While avoiding moving obstacles, a new travel route is planned. When the inspection unmanned vehicle travels according to the planned new travel route, it can continue to scan moving obstacles to plan a travel route to avoid obstacles. The new travel route generated by the obstacle avoidance of the inspection unmanned vehicle is the trajectory space.

7. An improved method for inspection unmanned vehicles based on dynamic path planning according to claim 1 or 6, characterized in that: The evaluation of the trajectory in the generated trajectory space in step 4 is specifically carried out as follows: determine whether the driving route of the inspection unmanned vehicle generated in the trajectory space will collide with moving obstacles. If a collision occurs, step 3 is executed again to generate the latest trajectory space; if no collision occurs, the trajectory with the shortest route and the best obstacle avoidance effect is selected.

8. An improved method for inspection unmanned vehicles based on dynamic path planning according to claim 7, characterized in that: The specific operation method for generating the optimal path in step 5 is as follows: evaluate and judge the optimal trajectory selected in step 4 to determine whether the optimal trajectory of the inspection unmanned vehicle has reached the target point. If it has reached the target point, generate the optimal path; otherwise, generate the trajectory space map and repeat steps 3 and 4 until the inspection unmanned vehicle safely reaches the target point.

Images