Bionic simulation plant phototropism global path planning method based on RRT algorithm

Abstract

The application proposes a global path search method of bionic plant phototropism on the basis of the original RRT algorithm, which optimizes the algorithm in the aspects of poor random expansion directionality, slow convergence speed, low search efficiency in narrow areas, and zigzag path planning. The bionic design in the application simulates the phototropism of plants, so that when the random tree selects and expands nodes on a series of concentric circles with the target point as the center, the speed of the random tree expanding towards the target point is accelerated, and the convergence efficiency of the algorithm is improved. In order to improve the safety during the expansion of the random tree towards the target point, the obstacles are processed, the obstacles with close distances are combined, and the obstacles are inflated to improve the safety. Secondly, the adaptive step length is designed, and the random tree changes the step length according to the density of the obstacles to better adapt to the complex environment. After extracting the key points of the generated path, the curve of the path is optimized to make the path smooth and meet the driving requirements of the intelligent vehicle. Finally, MATLAB is used for simulation analysis to verify the effectiveness of the algorithm.

Description

Technical Field

[0001] This invention belongs to the field of global path planning technology for intelligent driving vehicles and mobile robots, specifically relating to a biomimetic global path planning method for simulating the phototropism of plants based on the RRT algorithm. Background Technology

[0002] With the development of artificial intelligence and 5G technology, the world has ushered in a wave of unmanned and intelligent development, with unmanned factories, driverless cars, and intelligent robots increasingly appearing in our lives. Cars, as an important means of transportation, are also becoming increasingly electronic and intelligent. Building smart transportation systems is beneficial for alleviating traffic congestion, improving commuting efficiency, and reducing traffic accidents. Intelligent driving technology mainly includes environmental perception, navigation and positioning, path planning, and control decision-making. Path planning is a crucial component of intelligent driving and is of great significance to the development of intelligent driving technology.

[0003] Currently, research institutions and automakers worldwide are actively engaged in the development of intelligent driving technology, with path planning being an indispensable component. The effectiveness of path planning directly impacts the safety and comfort of intelligent vehicles. Path planning refers to the algorithmic planning of a safe, feasible, and collision-free path within a known environment. Path planning can be divided into global planning and local planning. Global planning utilizes known map information, static obstacle information, and start and end point information to plan a feasible path on the map that avoids static obstacles. Local path planning primarily targets dynamic obstacle avoidance, employing sensors such as LiDAR and cameras to perceive dynamic obstacles in the environment and perform real-time obstacle avoidance.

[0004] The Path Response Time (RRT) algorithm is a classic path planning algorithm that has been extensively studied and is often used in path planning for robots, drones, and autonomous vehicles. Summary of the Invention

[0005] This invention proposes an improvement upon the traditional RRT algorithm, resulting in higher planning efficiency and safer, more reasonable paths in static environments. The main design goal is to efficiently plan a safe, collision-free feasible path on a known map. Based on an improved global path planning method for the RRT algorithm, it optimizes aspects such as poor guidance during random expansion, slow convergence speed, low search efficiency in narrow areas, and tortuous planned paths.

[0006] The work performed in this invention includes at least the following: First, summarizing the feasibility and shortcomings of previous RRT algorithms, and proposing a biomimetic RRT algorithm (B-RRT, Bionics-Rapidly-exploring Random Tree) to address the problems of poor guidance and slow convergence speed in random expansion of RRT algorithms. This algorithm can improve path planning efficiency and reduce path tortuosity. Second, preprocessing obstacles in the environment by merging and expanding nearby obstacles to avoid sampling points being too close to obstacles, thereby improving path safety. Third, extracting key points in the random tree path and optimizing path reconnection and curves. Finally, simulation experiments using MATLAB are conducted to verify the effectiveness of the algorithm.

[0007] The specific technical solution adopted by this invention to solve its technical problem is as follows:

[0008] A biomimetic global path planning method for simulating plant phototropism based on the RRT algorithm is characterized by the following steps:

[0009] Step 1: Obtain obstacle coordinates, intelligent vehicle kinematics model, map information, initialize location information, and obtain target point information;

[0010] Step 2: Pre-process obstacles in the environment, merging and expanding obstacles that are close together;

[0011] Step 3: Obtain sampling points and determine their feasibility;

[0012] Step 4: Determine the density of obstacles and expand the random tree using a variable step size RRT based on safety assessment;

[0013] Step 5: Expand the random tree and determine if the target point has been reached. If not, return to Step 3; if so, proceed to the next step.

[0014] Step 6: Connect the feasible paths from the target point to the destination, extract the key points in the path, and reconnect the paths;

[0015] Step 7: Path optimization, path smoothing, and completion of path planning.

[0016] Furthermore, in step one, the intelligent vehicle uses cameras, LiDAR, IMU, and GPS to create an environmental map or directly obtain map information.

[0017] Initialize the intelligent vehicle's own position and obstacle information on the map. Since static obstacle information is known in the map, it is treated as a rectangle. Let the maximum and minimum x and y coordinates of a given obstacle be x, respectively. max x min y max ymin Then the coordinates of its four vertices are A(x) min ,y min ), B(x) max ,y min ), C(x) max ,y max ), D(x min ,y max A simplified kinematic model of the vehicle is established, and kinematic constraints are integrated into the path planning process to ensure that the planned path meets the vehicle's driving requirements.

[0018] This leads to the kinematic equations of the Ackermann steering driverless car.

[0019] Further, in step two, the environment is preprocessed. Let the length of the unmanned vehicle be 'a' and the width be 'b'. Let all obstacles be rectangular. If the distance between two adjacent obstacles is less than or equal to S1, then the gap is filled and they are merged into the same obstacle. Areas where the distance between obstacle surfaces is less than S2 are designated as infeasible areas, where new nodes cannot be generated, i.e., these areas are considered impassable. Areas where the distance is greater than or equal to S2 and less than or equal to S3 are designated as restricted areas, where new nodes have a certain probability of being generated, i.e., these areas are considered finitely passable areas. The formulas for S1, S2, and S3 are as follows:

[0020] S1=k·b

[0021] Where k is a safety factor greater than 1;

[0022]

[0023]

[0024] The probability P of obtaining sampling points in each region is given by the following formula, where p o The probability is set, and k is the distance between the sampling point and the obstacle:

[0025]

[0026] Furthermore, in step three, B-RRT is used to mimic the characteristics of plant growth, with the target point analogous to a light source and sparse obstacle areas analogous to open areas. Random nodes are selected on a series of concentric arcs centered on the target point. The formula for selecting random nodes is as follows:

[0027]

[0028] Sampling radius r:

[0029]

[0030] Where, q rand (x,y) are the coordinates of the random point to be selected; n1 and n2 are independent random numbers between 0 and 1; These are the x and y coordinates of the target point, respectively. R is the farthest distance between the two points on the map. If it is a rectangular map, it is the distance along the diagonal of the rectangle. m is a coefficient of 1 to 2.

[0031] Since the random points sampled by the random node selection formula may appear outside the map, it is necessary to remove the sampling points that appear outside the map before resampling.

[0032] Furthermore, in step four, the variable step size RRT based on safety assessment increases the step size of random tree expansion when fewer obstacles are detected in the surrounding environment, and the step size is determined according to the number of obstacles in the surrounding environment.

[0033] Through multiple explorations, the number of collisions between the exploration node and the nearest node and with obstacles is recorded. The density of obstacles in the range is determined based on the number of collisions, and then the step size is adjusted accordingly.

[0034] Formula for selecting random sampling points:

[0035]

[0036] Where, q rand (x, y) represents the coordinates of a random sampling node, and n1 and n2 are uncorrelated random numbers between 0 and 1. max Y max These represent the length and width of the map, respectively; n is a random number between 0 and 1, distinct from n1 and n2; P is the probability value for selecting the target point as a sampling node; and q... goal (x,y) are the coordinates of the target point.

[0037] Furthermore, in step five, it is determined whether the distance between the current sampling point and the target point is within the specified threshold. If the target point and the sampling point are directly connected, the path search is completed; otherwise, return to step three to expand the random tree again.

[0038] Furthermore, in step six, let the extracted path points be {P}. k Given a string of nodes k = 1, 2, ..., n, starting from node P1, connect nodes P1 and P3. If line segment P1P3 does not collide with an obstacle (i.e., the distance from line segment P1P3 to the obstacle is greater than a preset threshold), then continue connecting P1 and P4. If line segment P1P4 does not collide with an obstacle, then continue connecting P1 and P5, and so on, until P1 and P3 are connected. k And line segment P1P k After passing the obstacle, P will k-1Connect to P1, remove redundant intermediate points, and update the path; from node P k-1 Repeat the above steps until there are no redundant turning points in the path. Record these path nodes as the path's key points.

[0039] Furthermore, in step seven, Bézier curves are used for path optimization.

[0040] This invention and its preferred scheme propose a biomimetic global path search method based on the phototropism of plants, optimizing aspects such as poor guidance during random expansion, slow convergence speed, low search efficiency in narrow areas, and tortuous planned paths. The biomimetic design in this invention simulates the phototropism of plants, allowing the random tree to select expansion nodes on a series of concentric circles centered on the target point, thus accelerating the expansion speed of the random tree towards the target point and improving the algorithm's convergence efficiency. During the expansion process, obstacles are addressed to improve safety by merging closely spaced obstacles and expanding existing obstacles. Secondly, an adaptive step size is designed, where the random tree adjusts the step size according to the density of obstacles to better adapt to complex environments. After extracting key points from the generated path, curve optimization is performed to smooth the path and meet the driving requirements of intelligent vehicles. Finally, simulation analysis using MATLAB verifies the effectiveness of the algorithm. Attached Figure Description

[0041] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments:

[0042] Figure 1 This is a schematic diagram of obstacle coordinates according to an embodiment of the present invention;

[0043] Figure 2 A schematic diagram of the two-wheeled kinematic model of the unmanned vehicle according to an embodiment of the present invention;

[0044] Figure 3 Schematic diagram of environmental map obstacle preprocessing in an embodiment of the present invention;

[0045] Figure 4 Schematic diagram of a variable step size exploration node according to an embodiment of the present invention;

[0046] Figure 5 Schematic diagram comparing algorithm simulations in embodiments of the present invention;

[0047] Figure 6 A schematic diagram comparing the simulation time of the algorithm in this embodiment of the invention;

[0048] Figure 7 A schematic diagram illustrating the determination of whether the target point has been reached according to an embodiment of the present invention;

[0049] Figure 8 A schematic diagram of local target point selection in an embodiment of the present invention;

[0050] Figure 9 A schematic diagram of global path planning and path optimization using the B-RRT algorithm in this embodiment of the invention;

[0051] Figure 10 Flowchart of the method path planning in this embodiment of the invention. Detailed Implementation

[0052] To make the features and advantages of this patent more apparent and understandable, specific embodiments are provided below for detailed explanation:

[0053] It should be noted that the following detailed descriptions are illustrative and intended to provide further explanation of this application. Unless otherwise specified, all technical and scientific terms used in this specification have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0054] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this application. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0055] The present invention will be further described below through a specific embodiment, including the following steps:

[0056] Step 1: Obtain obstacle coordinates, intelligent vehicle kinematics model, map information, initialize location information, and target point information;

[0057] Intelligent vehicles use sensors such as cameras, LiDAR, IMU, and GPS to create environmental maps or directly acquire map information. They initialize their own position and obstacle information on the map. In a known map, static obstacle information is known; these are considered as rectangles. Let the maximum and minimum x and y coordinates of a given obstacle be x, respectively. max x min y max y min Then the coordinates of its four vertices are A(x) min ,y min ), B(x) max ,y min ), C(x) max ,y max ), D(x min ,y max ),like Figure 1 As shown. Next, a vehicle kinematic model is established and simplified as follows. Figure 2As shown. Kinematic constraints are incorporated into the path planning process to ensure that the planned path meets the driving requirements of the vehicle. The following kinematic model and its formulas will be explained in detail.

[0058] Among them, (x f ,y f ), (x r ,y r These are the center positions of the front and rear axles, which are the simplified positions of the front and rear wheels (hereinafter referred to as front wheels and rear wheels). f v r These are the linear velocities of the front and rear wheels, respectively, and δ is the steering angle of the front wheel. is the heading angle of the autonomous vehicle, l is the wheelbase of the autonomous vehicle, and R is the turning radius of the autonomous vehicle.

[0059] Based on orthogonal decomposition, the velocity relationship is as follows:

[0060]

[0061] From geometric relationships, the yaw rate can be obtained as:

[0062]

[0063] The kinematic equations of the Ackermann steering autonomous vehicle are obtained by integration:

[0064]

[0065] Step 2: Preprocess obstacles in the environment, and merge and expand obstacles that are close together;

[0066] The environment is preprocessed, especially the obstacles. Let the length of the autonomous vehicle be *a* and the width be *b*. Assume all obstacles are rectangular. If the distance between two adjacent obstacles is less than or equal to *S1*, then fill the gap and merge them into a single obstacle. Figure 3 As shown in (a), to avoid the random tree sampling points being too close to obstacles, thus increasing the risk factor for the intelligent vehicle, the obstacles are expanded. Areas with a distance less than S2 from the obstacle surface are designated as infeasible areas, where new nodes cannot be generated; these areas are considered impassable. Areas with a distance greater than or equal to S2 and less than or equal to S3 are designated as restricted areas, where new nodes have a certain probability of being generated; these areas are considered finitely passable areas. A schematic diagram is shown below. Figure 3 As shown in (b), the formulas for S1, S2, and S3 are as follows:

[0067] S1=k·b

[0068] Where k is a safety factor greater than 1.

[0069]

[0070]

[0071] The probability P of obtaining sampling points in each region is given by the following formula, where p o The probability is set manually, and k is the distance between the sampling point and the obstacle. Based on this formula, further improvements to path safety can be made.

[0072]

[0073] Step 3: Obtain sampling points and determine their availability;

[0074] The B-RRT proposed in this invention mimics the characteristics of plant growth, analogizing the target point to a light source and sparse areas of obstacles to open spaces. Furthermore, when selecting random nodes, they are positioned on a series of concentric arcs centered on the target point. The formula for selecting random nodes in the B-RRT algorithm is as follows.

[0075]

[0076] In the above formula, the sampling radius r is:

[0077]

[0078] Where, q rand (x,y) are the coordinates of the random point to be selected; n1 and n2 are independent random numbers between 0 and 1; Here, x and y are the x and y coordinates of the target point, respectively; R is the furthest distance between the two points on the map; for a rectangular map, it is the distance along the diagonal of the rectangle; and m is a coefficient between 1 and 2. Random points sampled using this formula may appear outside the map, therefore, sampling points that do not meet the planning requirements need to be removed before resampling.

[0079] Step 4: Determine the density of obstacles and select an appropriate step size to expand the random tree;

[0080] This invention proposes a variable step size RRT based on relevant concepts and safety assessment, such as... Figure 4 As shown, when there are few obstacles in the surrounding environment, the step size of the random tree expansion can be appropriately increased. The optimization and improvement of the step size is determined according to the number of obstacles in the surrounding environment, which is the variable step size RRT. The algorithm will be explained in detail below. Figure 4The small black circles represent extended tree nodes, black rectangles represent obstacles, triangles represent target points, and the area enclosed by the dashed box represents the exploration range centered on the nearest sampling node (not shown in the figure). White dots represent randomly explored nodes. Through multiple explorations, the number of collisions between the exploration node and the nearest node, as well as with obstacles, is recorded. The density of obstacles in this range is qualitatively determined based on the number of collisions, and the step size is adjusted accordingly.

[0081] When using the Variable Step Time (RRT) for path planning, the random tree expansion step size changes constantly. In this simulation experiment, for simplicity, the random tree expansion step size was chosen to vary between 1, 1.5, and 2. The main idea is to use a smaller step size in areas with dense obstacles to avoid discarding new nodes due to collisions with obstacles; and a larger step size in areas with sparse obstacles to accelerate the expansion of the random tree and improve algorithm efficiency. The density of obstacles is determined by exploring a 2m radius around the nearest node, for a total of ten explorations. The number of times no obstacle collision occurs is recorded. If the number is greater than 6, the step size is 2m; if it is greater than 3 but less than or equal to 6, the step size is 1.5m; and if it is less than or equal to 3, the step size is 1m. The formula is as follows:

[0082]

[0083] Where B is the expansion step size of the random tree, and n is the number of collisions explored around the nearest node.

[0084] In research aimed at improving algorithm efficiency and accelerating the expansion of random trees towards the target point, many researchers have proposed target-biased strategies. Broadly speaking, this means that during expansion, the distribution of sampling points is not uniformly selected across the entire environment, but rather the probability of selecting sampling points in a specific region is artificially increased, thus accelerating the expansion of the random tree towards the target point. In the algorithm itself, target bias specifically refers to a certain probability of selecting the target point as a sampling point, allowing the random tree to directly expand towards the target point and accelerating the algorithm's convergence speed. The original RRT algorithm's formula for selecting random sampling points is as follows:

[0085] q rand (x,y)=(n1×X max ,n2×Y max )

[0086] Where, q rand (x, y) represents the coordinates of a random sampling node, and n1 and n2 are uncorrelated random numbers between 0 and 1. max Y max These are the length and width of the map, respectively.

[0087] The improved target-biased RRT algorithm in this embodiment uses the following formula for selecting random sampling points:

[0088]

[0089] Where n is a random number between 0 and 1, distinct from n1 and n2, P is the probability value of selecting the target point as the sampling node, and in this simulation experiment, P is taken as 0.2, q goal (x,y) are the coordinates of the target point, and the other symbols have the same meaning as above.

[0090] To compare the target bias RRT algorithm and the B-RRT algorithm, this embodiment conducts simulation experiments on both. Figure 5 The simulation comparison chart clearly shows that the target-biased RRT algorithm has sparser sampling points. In the later stages of path planning, the algorithm can rely on a certain probability to select the target point as the direction for the expansion of the random tree, solving the problem of the path planning lingering near the target point. In contrast, the random sampling points of the B-RRT algorithm are densely distributed in areas close to the target point and sparsely distributed in areas far from the target point. This sampling characteristic also allows the random tree to preferentially expand towards the target point. To clearly demonstrate the difference between the target-biased RRT algorithm and the B-RRT algorithm, each algorithm was simulated 100 times, and the time taken was recorded, rounded to two decimal places. Figure 6 This is a comparison chart showing the time required for path planning by the two RRT algorithms.

[0091] Simulation data shows that the average time of the target bias RRT is 0.89, with a variance of 0.04196, while the average time of the B-RRT is 0.53, with a variance of 0.0088. This indicates that the B-RRT time is 40% lower than the target bias RRT time. In addition, the stability of the B-RRT algorithm is improved to a certain extent, which significantly improves the algorithm efficiency.

[0092] Step 5: Expand the random tree and determine if the target point has been reached. If the target point has not been reached, return to Step 3; if it has been reached, proceed to the next step.

[0093] Determine if the distance between the current sampling point and the target point is within a specified threshold. If the target point and the sampling point are directly connected, the path search is complete. Figure 7 As shown; otherwise, return to step three to expand the random tree again.

[0094] Step 6: Connect the feasible paths from the target point to the destination, extract the key points in the path, and reconnect the paths;

[0095] In this step, the B-RRT algorithm has already planned the global path, and the random tree connection is used to connect the starting point to the target point, such as... Figure 8 Blue extended tree; then we need to extract the path nodes in this global path, let the extracted path nodes be {P}. kGiven a string of nodes k = 1, 2, ..., n, starting from node P1, connect nodes P1 and P3. If line segment P1P3 does not collide with an obstacle (i.e., the distance from line segment P1P3 to the obstacle is greater than a preset threshold), then continue connecting P1 and P4. If line segment P1P4 does not collide with an obstacle, then continue connecting P1 and P5, and so on, until P1 and P3 are connected. k And line segment P1P k After passing the obstacle, P will k-1 Connect to P1, remove redundant intermediate points, and update the path; from node P k-1 Repeat the above steps until there are no redundant turning points in the path. Record these path nodes as the path's key points.

[0096] Figure 8 The sampling points have been hidden, and only the original blue planned path is retained. The green path is the path retained based on key points, the red asterisks * indicate the extracted key points of the path, and the red pentagrams represent the global target points. It can be seen that the final retained path has a significantly reduced total length compared to the initial path, and the number of turning points has been greatly reduced. The original path had 49 turning points, while after removing redundant intermediate points, the final path has only 7 turning points, a reduction of 86%. However, although the number of turning points has been reduced, the path is not smooth enough to meet the driving requirements of intelligent vehicles. The next step will be to optimize the green path using curves.

[0097] Step 7: Path optimization, path smoothing, completion of path planning, and visualization of the global path in MATLAB;

[0098] The previous step extracted nodes of the global path and further extracted key points of the global path according to rules, which optimized the path to some extent. However, the reconnected paths lacked curve optimization and were not smooth enough to meet the driving requirements of intelligent vehicles. Therefore, Bézier curves were introduced for path optimization. The path optimized by Bézier curves is smooth and meets the driving requirements of the vehicle. In the path optimization process, not only were key points of the global path considered to ensure that the path can avoid obstacles, but dynamic constraints were also added to take into account the vehicle's steering, acceleration, etc. Only the path after curve optimization truly meets the driving requirements of the vehicle. Figure 9 (b) is the global path optimized by Bézier curves, visualized in MATLAB. Figure 9 In (a), the black curve represents the original global path, and the dashed line represents the path formed after extracting path keys. The comparison between the left and right sides shows that the optimized path closely matches the red global path to a certain extent, while also considering vehicle movement and obstacle avoidance, especially avoiding sharp turns with large curvatures. Therefore, the improved algorithm of this invention has good practical value in optimizing its path. Its overall process is as follows: Figure 10 As shown.

[0099] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

[0100] This patent is not limited to the above-described preferred embodiments. Anyone can derive other forms of biomimetic simulated plant phototropism global path planning methods based on the RRT algorithm under the guidance of this patent. All equivalent changes and modifications made within the scope of this patent application shall fall within the scope of this patent.

Claims

1. A biomimetic global path planning method for simulating plant phototropism based on the RRT algorithm, characterized in that, Includes the following steps: Step 1: Obtain obstacle coordinates, intelligent vehicle kinematics model, map information, initialize location information, and obtain target point information; Step 2: Pre-process obstacles in the environment, merging and expanding obstacles that are close together; Step 3: Obtain sampling points and determine their feasibility; In step three, B-RRT is used to mimic the characteristics of plant growth, with the target point analogous to a light source and sparse obstacle areas analogous to open areas. Random nodes are selected on a series of concentric arcs centered on the target point. The formula for selecting random nodes is as follows: Sampling radius : in, The coordinates of the random point to be selected; , These are independent random numbers between 0 and 1; , These are the x and y coordinates of the target point, respectively. This represents the furthest distance between two points on the map. If it's a rectangular map, it's the distance along the diagonal of the rectangle. A coefficient between 1 and 2; Since the random points sampled by the random node selection formula may appear outside the map, it is necessary to remove the sampling points that appear outside the map before resampling. Step 4: Determine the density of obstacles and expand the random tree using a variable step size RRT based on safety assessment; In step four, the variable step size RRT based on safety assessment increases the step size of random tree expansion when fewer obstacles are detected in the surrounding environment, and the step size is determined according to the number of obstacles in the surrounding environment. Through multiple explorations, the number of collisions between the exploration node and the nearest node and the obstacle is recorded. Based on the number of collisions, the density of obstacles in the exploration range centered on the nearest point to the sampling point is qualitatively determined, and then the step size is adjusted. Formula for selecting random sampling points: in, The coordinates of the randomly sampled nodes, , These are unrelated random numbers between 0 and 1. , These are the length and width of the map, respectively. It is different from , A random number between 0 and 1. It is the probability value of selecting the target point as the sampling node. These are the coordinates of the target point; Step 5: Expand the random tree and determine if the target point has been reached. If not, return to Step 3; if so, proceed to the next step. Step 6: Connect the feasible paths from the target point to the destination, extract the key points in the path, and reconnect the paths; Step 7: Optimize the path to smooth it out and complete the path planning.

2. The biomimetic simulated plant phototropism global path planning method based on RRT algorithm according to claim 1, characterized in that: In step one, the intelligent vehicle uses cameras, LiDAR, IMU, and GPS to create an environmental map or directly obtain map information. Initialize the intelligent vehicle's own position and obstacle information on the map. Since static obstacle information is known in the map, it is treated as a rectangle. Let the maximum and minimum x and y coordinates of a given obstacle be respectively... Then the coordinates of its four vertices are , A simplified vehicle kinematic model is established, and kinematic constraints are integrated into the path planning process to ensure that the planned path meets the vehicle's driving requirements. This leads to the kinematic equations of the Ackermann steering driverless car.

3. The biomimetic simulated plant phototropism global path planning method based on the RRT algorithm according to claim 1, characterized in that: In step two, the environment is preprocessed. Let the length of the unmanned vehicle be... , width is ; Assuming all obstacles are rectangular, if the distance between two adjacent obstacles is less than or equal to... Then, it fills the gap between two adjacent obstacles, merging the two adjacent obstacles into a single obstacle; if the distance between the surfaces of the obstacles is less than... A region is designated as an inaccessible region, meaning new nodes cannot be generated within it. In other words, the region is considered impassable, and nodes with a distance greater than or equal to [a specified value] are considered to be inaccessible. less than or equal to The area is set as a restricted access area, and new nodes have a certain probability of being generated in this area; that is, this area is considered a restricted access area. The formula is as follows: in, A safety factor greater than 1; The probability P of obtaining sampling points in each region is given by the following formula, where For the set probability, The distance between the sampling point and the obstacle: 。 4. The biomimetic simulated plant phototropism global path planning method based on RRT algorithm according to claim 1, characterized in that: In step five, it is determined whether the distance between the current sampling point and the target point is within the specified threshold. If so, the target point and the sampling point are directly connected, and the path search is completed; otherwise, return to step three to expand the random tree again.

5. The biomimetic simulated plant phototropism global path planning method based on RRT algorithm according to claim 1, characterized in that: In step six, let the extracted path points be... From node Begin by connecting nodes. If line segment The line segment does not collide with obstacles. If the distance to the obstacle is greater than a preset threshold, the connection continues. If line segment If no collision occurs with an obstacle, the connection continues. And so on, until a connection is made. And line segment After passing through the obstacle, and Connect them, remove redundant intermediate points, and update the path; from the nodes Repeat the above steps until there are no redundant turning points in the path. Record these path nodes as the path's key points.

6. The biomimetic simulated plant phototropism global path planning method based on RRT algorithm according to claim 1, characterized in that: In step seven, Bézier curves are used for path optimization.

Images