An unmanned tracked vehicle path planning method and system based on a motion dynamics RRT improved algorithm

Abstract

The application provides an unmanned tracked vehicle path planning method and system based on a motion dynamics RRT improved algorithm, and belongs to the unmanned tracked vehicle path planning field. In order to solve the problem that the existing algorithm planning path produces a convolution, causes the path to be not smooth enough or even to fail, and increases the calculation complexity and planning time cost, the application obtains map information needing to be planned, initializes parameters and a path tree; a target bias strategy is used to randomly sample a shape point, a nearest neighbor node of the random shape on the path tree is searched; the distance from the node to a target position is calculated, a control set proportion factor is calculated; at the nearest neighbor node, a control set is selected and a control vector is randomly sampled, and then the proportion factor is applied to generate a final control vector; the control vector is input into a tracked vehicle kinematics model to be integrated to generate a new shape; collision detection is performed; the cycle is continued until the new shape enters a target neighborhood, and finally a parent node chain is backtracked to obtain a path.

Description

Technical Field

[0001] This invention relates to the field of unmanned tracked vehicle path planning technology, and more specifically, to an unmanned tracked vehicle path planning method and system based on an improved RRT algorithm of motion dynamics. Background Technology

[0002] Path planning technology, as a fundamental and crucial topic in the research of unmanned tracked vehicles, aims to find a path in the configuration space C starting from the initial configuration q. init C free To the target configuration q goal C free A collision-free path that satisfies specific constraints. Sampling-based motion planning does not require explicitly obtaining C. free Configuration space allows for significant computational savings, making it an effective method for obtaining paths. A representative sampling-based motion planning algorithm is the Rapidly-exploring Random Tree (RRT) algorithm. The RRT algorithm initializes the tree from the starting point. It then iteratively executes the following steps: randomly sampling in the configuration space C; searching the tree for the node closest to the sampling point; extending towards the sampling point by a fixed step size to generate a new node and performing collision detection. If no collision occurs, the new node is added to the tree. This process continues until the new node approaches the target area, and finally, the planned path is obtained by backtracking from the target node back to the starting point. However, the RRT algorithm expands new nodes using a fixed-step linear growth method, which ignores the nonholonomic constraints of the unmanned tracked vehicle system. This may result in the actual path deviating from the planned path, leading to path failure. Therefore, it is necessary to consider the nonholonomic constraints of the unmanned tracked vehicle in motion planning.

[0003] To overcome the limitation of the RRT algorithm in ignoring the nonholonomic constraints of the planning system and improve the success rate of path planning, scholars at home and abroad have conducted a great deal of research on improvements. Based on the RRT algorithm, some researchers have proposed the Kinodynamic RRT algorithm. This algorithm considers the kinematic model of the planning system and obtains new nodes by integrating the kinematic state transition equations of the system at the nearest node. This node expansion method explicitly embeds the nonholonomic constraints of the planning system into the planning process, thus effectively solving the problem of nonholonomic constraints on the planning system. However, this algorithm does not restrict the control vector during planning, which can lead to looping in the planned path, resulting in an uneven path or even failure. Furthermore, the algorithm requires a large amount of sampling during planning, which increases computational complexity and planning time costs, making it unsuitable for real-time use in unmanned tracked vehicles.

[0004] In summary, it can be seen that existing improvements to the dynamics RRT still have some defects and shortcomings, which affect their application in autonomous navigation of unmanned tracked vehicles. Summary of the Invention

[0005] The technical problem to be solved by this invention is:

[0006] To address the issues that existing motion dynamics RRT algorithms do not impose restrictions on control vectors, leading to looping in the planned path, resulting in an uneven path or even failure, and that require extensive sampling during planning, increasing computational complexity and planning time costs.

[0007] The technical solution adopted by the present invention to solve the above-mentioned technical problems is as follows:

[0008] This invention provides a path planning method for unmanned tracked vehicles based on an improved RRT algorithm of kinematics, comprising the following steps:

[0009] S100. Obtain the environmental spatial map information for path planning, including map size, shape and position of obstacles, define the starting configuration and target configuration of the unmanned tracked vehicle, set the target neighborhood radius for terminating path exploration, calculate the Euclidean distance between the position components in the starting configuration and the target configuration, and initialize the path tree.

[0010] S200. Given the target bias probability, use the target bias strategy according to this probability to sample and generate random configurations in the map where path planning is to be performed.

[0011] S300: Find the closest random configuration obtained from sampling to the configuration on the path tree;

[0012] S400: Calculate the distance between the positional components in the nearest configuration and the endpoint configuration obtained in step S300, and calculate the control set scaling factor based on the distance.

[0013] S500. At the nearest configuration, random sampling is performed in the control set and the control set scaling factor is obtained according to step S400 to generate the final control vector. Then, the new configuration is generated by integrating the tracked vehicle state transition equation within a given time.

[0014] S600 Collision detection: Check if there is an obstacle between the new configuration and the nearest configuration on the path tree; if there is no collision, add the new node to the path tree; if a collision occurs, return to step S200 and resample.

[0015] S700. Determine whether the new configuration is located in the target neighborhood of the endpoint configuration. If so, it is considered that a path has been found, the loop is exited, and the planned path is obtained by backtracking from the target configuration along the parent node configuration. If not, repeat the above steps until a path is found.

[0016] Further, in step S100, the Euclidean distance between the position components in the starting configuration and the target configuration is calculated:

[0017] (1)

[0018] In the formula, Let Euclidean distance be the positional components of the starting configuration and the target configuration; the configuration of a tracked vehicle is represented as... , and It refers to the position of the center of gravity of the tracked vehicle, that is, the positional component of the tracked vehicle's configuration. These are the angular components of the tracked vehicle configuration; and It is the position component of the target configuration of the tracked vehicle; and It is the position component of the starting configuration of the tracked vehicle.

[0019] Further, in step S200, the following are included:

[0020] S210, Set target bias probability During each sampling, a probability is generated in a uniform distribution between 0 and 1. ,when Generate random configurations of tracked vehicles within the map space, or select target configurations elsewhere. This serves as the configuration for this random sampling.

[0021] S220, when Generate random configurations within the map space. :

[0022] (2)

[0023] In the formula, For map space The range of components, For map space The range of components; and The positional components of the randomly generated tracked vehicle configuration; For random configuration angle direction components; To generate randomly according to a uniform distribution The values ​​between.

[0024] Furthermore, in step S300, when searching for the nearest configuration on the path tree using a random configuration, the distance metric uses the following distance function:

[0025] (3)

[0026] In the formula, This is a distance metric function used to measure distance. and The distance between the two configurations; and for Positional components of the configuration for angular components of the configuration; and for Positional components of the configuration for angular components of the configuration; for and The smaller one.

[0027] Further, in step S400, the following is included:

[0028] S410. After obtaining the configuration closest to the random configuration on the path tree through step S300, calculate the Euclidean distance between the positional components of the nearest configuration and the destination configuration. ;

[0029] S420, Euclidean distance between the positional components in the nearest and final configurations obtained through S410. Then, make a judgment Is it greater than the cutoff distance? If it is less than, then a control set scaling factor is given. and If it is greater than 1, then calculate the control set scaling factor:

[0030] (4)

[0031] In the formula, It is a control set Scale factor; It is a control set Scale factor; Indicates to Round down to the nearest integer.

[0032] Further, in step S500, the following are included:

[0033] S510, Provide the control set for tracked vehicles. Obtain its speed control subset and angular velocity control subset and speed control subset The non-negative portion serves as a new velocity control subset. Perform subsequent sampling;

[0034] S520: The control set scaling factor obtained through S420 is combined with the control set to obtain the final control vector by random sampling according to a uniform distribution.

[0035] (5)

[0036] In the formula, The control vector obtained by sampling, Its velocity component, Its angular velocity component;

[0037] S530, the control vector obtained through S520, is then integrated within a given time according to the tracked vehicle state transition equation to generate a new configuration:

[0038] (6)

[0039] In the formula, The new configuration obtained by integration; It is the integration time; and New configuration for tracked vehicles Positional components; For the angular direction components of the new configuration; and For the closest positioning of the tracked vehicle Positional components; The nearest configuration angle direction component;

[0040] set up The preset threshold parameter is used when the sampled data is obtained. Value at When the interval is within a certain range, the formula for calculating the new configuration is:

[0041] (7).

[0042] Further, in step S700, the target configuration neighborhood is represented as:

[0043] (8)

[0044] In the formula, The target configuration neighborhood range; The radius of the allowable positional error for the target configuration; The radius of the allowable angular error for the target configuration.

[0045] An unmanned tracked vehicle path planning system based on the improved RRT algorithm of kinematics is provided. The system has program modules corresponding to the above steps and executes the steps in the above-mentioned unmanned tracked vehicle path planning method based on the improved RRT algorithm of kinematics during runtime.

[0046] A computer-readable storage medium storing a computer program configured to, when invoked by a processor, implement steps of a path planning method for an unmanned tracked vehicle based on an improved RRT algorithm of kinematics.

[0047] Compared with the prior art, the beneficial effects of the present invention are:

[0048] This invention discloses a path planning method and system for unmanned tracked vehicles based on an improved kinematic dynamics RRT algorithm. The method includes: acquiring map information for planning and initializing parameters and a path tree; randomly sampling configuration points using a target bias strategy and searching for the nearest neighbor node in the path tree corresponding to the random configuration; calculating the distance from the node to the target location and adaptively calculating the control set scaling factor; selecting the control set at the nearest neighbor node, randomly sampling control vectors, and then applying the scaling factor to generate the final control vector; integrating the control vectors into the tracked vehicle's kinematic model to generate a new configuration; performing collision detection on the new configuration, and adding it to the path tree if safe; repeating this process until the new configuration enters the target neighborhood, and finally backtracking the parent node chain to obtain the path.

[0049] This invention provides a path planning method and system for unmanned tracked vehicles based on the kinematic dynamics RRT algorithm. It uses a target bias strategy to accelerate convergence and adaptively adjusts the control scaling factor by calculating the distance from the nearest node to the target, making the generated control variables more closely match the needs of the actual scenario, reducing the randomness of path exploration, and improving path planning efficiency. This invention embeds the kinematic model integral of the tracked vehicle into the path extension process, ensuring that the new configuration strictly conforms to the vehicle's motion characteristics, avoiding infeasible paths planned by traditional RRT algorithms, and providing motion constraint guarantees, thus improving the quality of path planning. This invention selects a suitable control set and combines it with adaptive adjustment of the control scaling factor to generate appropriate control vectors, reducing ineffective extensions. Simultaneously, it generates physically executable smooth paths through state integration, improving the success rate of path planning. Attached Figure Description

[0050] Figure 1 This is a flowchart of a path planning method for unmanned tracked vehicles based on an improved RRT algorithm in motion dynamics, as described in an embodiment of the present invention.

[0051] Figure 2 The image shows the planned trajectory of the traditional motion dynamics RRT algorithm in an obstacle-free path in this embodiment of the invention, where (a) is the path tree generated by exploration, and (b) is the overall planned path obtained from the path tree;

[0052] Figure 3The image shows the planned trajectory of the conventional motion dynamics RRT algorithm on a path with obstacles in an embodiment of the present invention, where (a) is the path tree generated by exploration, and (b) is the overall planned path obtained from the path tree;

[0053] Figure 4 The above is a trajectory diagram of the improved motion dynamics RRT algorithm in an obstacle-free path in an embodiment of the present invention, wherein (a) is a path tree diagram generated by exploration, and (b) is the overall planned path diagram obtained from the path tree;

[0054] Figure 5 The above is a trajectory diagram of the improved motion dynamics RRT algorithm in an obstacle-free path in an embodiment of the present invention, wherein (a) is a path tree diagram generated by exploration, and (b) is the overall planned path diagram obtained from the path tree. Detailed Implementation

[0055] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0056] Specific Implementation Plan 1: Combining Figure 1 As shown, this invention provides a path planning method for unmanned tracked vehicles based on an improved RRT algorithm of kinematics, comprising the following steps:

[0057] S100. Obtain the environmental spatial map information for path planning, including map size, shape and position of obstacles, define the starting configuration and target configuration of the unmanned tracked vehicle, set the target neighborhood radius for terminating path exploration, calculate the Euclidean distance between the position components in the starting configuration and the target configuration, and initialize the path tree.

[0058] The Euclidean distance between the positional components of the starting configuration and the target configuration is calculated using the following method:

[0059] (1)

[0060] In the formula, Let Euclidean distance be the positional components of the starting configuration and the target configuration; the configuration of a tracked vehicle is represented as... , and It refers to the position of the center of gravity of the tracked vehicle, that is, the positional component of the tracked vehicle's configuration. These are the angular components of the tracked vehicle configuration; and It is the position component of the target configuration of the tracked vehicle; and It is the position component of the starting configuration of the tracked vehicle;

[0061] S200. Given the target bias probability, use the target bias strategy according to this probability to sample and generate random configurations in the map where path planning is to be performed.

[0062] include,

[0063] S210, Set target bias probability (Set it according to your needs, for example, you can set it to...) =0.05), and at each sampling, a probability is generated in a uniform distribution between 0 and 1. ,when Generate random configurations of tracked vehicles within the map space, or select target configurations elsewhere. This serves as the configuration for this random sampling.

[0064] S220, when Generate random configurations within the map space. The specific implementation process is as follows:

[0065] (2)

[0066] In the formula, For map space The range of components, For map space The range of components; and The positional components of the randomly generated tracked vehicle configuration; For random configuration angle direction components; To generate randomly according to a uniform distribution The values ​​between;

[0067] S300: Find the closest random configuration obtained from sampling to the configuration on the path tree;

[0068] When finding the nearest configuration on a path tree for a random configuration, the distance metric used is the following distance function:

[0069] (3)

[0070] In the formula, This is a distance metric function used to measure distance. and The distance between the two configurations; and for Positional components of the configuration for angular components of the configuration; and for Positional components of the configuration for angular components of the configuration; for and The smaller one;

[0071] S400. Calculate the distance between the positional components in the nearest configuration and the endpoint configuration, and calculate the control set scaling factor according to a certain strategy based on the distance.

[0072] include,

[0073] S410. After obtaining the configuration closest to the random configuration on the path tree through step S300, calculate the Euclidean distance between the positional components of the nearest configuration and the destination configuration. ;

[0074] S420, Euclidean distance between the positional components in the nearest and final configurations obtained through S410. Then, make a judgment Is it greater than the cutoff distance? If it is less than, then a control set scaling factor is given. and If it is greater than 1, then calculate the control set scaling factor, specifically:

[0075] (4)

[0076] In the formula, It is a control set Scale factor; It is a control set Scale factor; Indicates to Round down;

[0077] S500. At the nearest configuration, random sampling is performed in the control set and the final control vector is generated according to the control set scaling factor. Then, the new configuration is generated by integrating the tracked vehicle state transition equation within a given time.

[0078] include,

[0079] S510, Provide the control set for tracked vehicles. Obtain its speed control subset and angular velocity control subset and speed control subset The non-negative portion serves as a new velocity control subset. Perform subsequent sampling;

[0080] S520: The control set scaling factor obtained through S420 is combined with the control set randomly sampled according to a uniform distribution to finally obtain the control vector. The formula for calculating the control vector is:

[0081] (5)

[0082] In the formula, The control vector obtained by sampling, Its velocity component, Its angular velocity component;

[0083] S530, the control vector obtained through S520, is then integrated within a given time according to the tracked vehicle state transition equation to generate a new configuration. The formula for calculating the new configuration is:

[0084] (6)

[0085] In the formula, The new configuration obtained by integration; It is a certain integration time, which can be set as needed, for example, it can be set to 4; and New configuration for tracked vehicles Positional components; For the angular direction components of the new configuration; and For the closest positioning of the tracked vehicle Positional components; The nearest configuration angle direction component;

[0086] set up The preset threshold parameter is used when the sampled data is obtained. Value at When the interval is within a certain range, the formula for calculating the new configuration is:

[0087] (7)

[0088] in You can set it according to your needs; for example, you can set it to 0.01.

[0089] S600 Collision detection: Check if there is an obstacle between the new configuration and the nearest configuration on the path tree; if there is no collision, add the new node to the path tree; if a collision occurs, return to step S200 and resample.

[0090] S700. Determine whether the new configuration is located in the target neighborhood of the endpoint configuration. If so, it is considered that a path has been found, the loop is broken, and the planned path is obtained by backtracking from the target configuration along the parent node configuration.

[0091] The target configuration neighborhood is represented as:

[0092] (8)

[0093] In the formula, The target configuration neighborhood range; The allowable positional error radius for the target configuration can be set according to requirements, for example, it can be set to 0.25; Set the allowable angular error radius for the target configuration. This can be set according to your needs; for example, it can be set to 0.3.

[0094] Specific implementation scheme two: The present invention provides an unmanned tracked vehicle path planning system based on the improved RRT algorithm of kinematics. The system has program modules corresponding to the above steps, and executes the steps in the above-mentioned unmanned tracked vehicle path planning method based on the improved RRT algorithm of kinematics during runtime.

[0095] The other combinations and connections in this implementation scheme are the same as in Specific Implementation Scheme 1.

[0096] Specific Implementation Scheme 3: The present invention provides a computer-readable storage medium storing a computer program configured to implement, when called by a processor, the steps of a path planning method for an unmanned tracked vehicle based on an improved RRT algorithm of kinematics.

[0097] The other combinations and connections in this implementation scheme are the same as in Specific Implementation Scheme 1.

[0098] Simulation Experiment

[0099] Based on two environmental maps, one without obstacles and the other with obstacles, the path planning performance of the improved motion dynamics RRT in this embodiment is compared with that of the traditional motion dynamics RRT. Figures 2-5 As shown. Figures 2-3 The path planned by the traditional RRT method of motion dynamics, wherein, Figure 2 The environmental spatial map has no obstacles. Figure 3 The environmental spatial map contains obstacles. Figure 2 and Figure 3 Subgraph (a) is the path tree generated during the exploration. Figure 2 and Figure 3 Subgraph (b) is the overall planning path obtained from the path tree. Figures 4-5 To improve the path planning of the RRT method in motion dynamics, where, Figure 4 The environmental spatial map has no obstacles. Figure 5 The environmental spatial map contains obstacles. Figure 4 and Figure 5 Subgraph (a) is the path tree generated during the exploration. Figure 4 and Figure 5 Subgraph (b) represents the overall planned path obtained from the path tree. Starting point configuration. The target configuration is in the lower left corner. In the upper right corner. As can be seen from the figure, the traditional motion dynamics RRT algorithm generates a rather tortuous path with a spiraling effect. The improved motion dynamics RRT algorithm can eliminate the spiraling effect, reduce unnecessary turns or path length, and thus generate a smooth path. It has a certain degree of guidance when exploring space, effectively avoiding inefficient spatial search, which performs better in complex environments. The improved motion dynamics RRT algorithm can flexibly adjust algorithm parameters and structure to meet different application scenarios and complexities.

[0100] Table 1 Path Parameters

[0101]

[0102] To avoid the randomness of the above results, 10 repeated experiments were conducted on both the traditional motion dynamics RRT and the improved motion dynamics algorithms in environments with and without obstacles. The average value was taken as the experimental evaluation index, and the results are shown in Table 2 below:

[0103] Table 2 Parameters of Repeated Experiment Path

[0104]

[0105] As shown in the table, for obstacle-free conditions, the improved motion dynamics RRT algorithm finds 43.4% fewer path tree nodes than the traditional motion dynamics RRT algorithm; for obstacle-affected conditions, the improved motion dynamics RRT algorithm finds 54.8% fewer path tree nodes than the traditional motion dynamics RRT algorithm, indicating a significant improvement in planning efficiency. In summary, the improved motion dynamics RRT path planning method proposed in this invention provides a new solution for path planning of unmanned tracked vehicles.

[0106] While the present invention has been disclosed above, its scope of protection is not limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention, and all such changes and modifications will fall within the scope of protection of the present invention.

Claims

1. A path planning method for unmanned tracked vehicles based on an improved RRT algorithm of kinematics, characterized in that, Includes the following steps: S100. Obtain the environmental spatial map information for path planning, including map size, shape and position of obstacles, define the starting configuration and target configuration of the unmanned tracked vehicle, set the target neighborhood radius for terminating path exploration, calculate the Euclidean distance between the position components in the starting configuration and the target configuration, and initialize the path tree. S200. Given the target bias probability, use the target bias strategy according to this probability to sample and generate random configurations in the map where path planning is to be performed. S300: Find the closest random configuration obtained from sampling to the configuration on the path tree; S400: Calculate the distance between the positional components in the nearest configuration and the endpoint configuration obtained in step S300, and calculate the control set scaling factor based on the distance. S500. At the nearest configuration, random sampling is performed in the control set and the control set scaling factor is obtained according to step S400 to generate the final control vector. Then, the new configuration is generated by integrating the tracked vehicle state transition equation within a given time. S600 Collision Detection: Checks whether there are obstacles between the new configuration and the nearest configuration on the path tree; If there is no collision, add the new node to the path tree; If a collision occurs, return to step S200 and resample; S700. Determine whether the new configuration is located in the target neighborhood of the endpoint configuration. If so, it is considered that a path has been found, the loop is exited, and the planned path is obtained by backtracking from the target configuration along the parent node configuration. If not, repeat the above steps until a path is found.

2. The path planning method for unmanned tracked vehicles based on the improved RRT algorithm of kinematics as described in claim 1, characterized in that: In step S100, the Euclidean distance between the position components in the starting configuration and the target configuration is calculated: （1） In the formula, Let Euclidean distance be the positional components of the starting configuration and the target configuration; the configuration of a tracked vehicle is represented as... , and It refers to the position of the center of gravity of the tracked vehicle, that is, the positional component of the tracked vehicle's configuration. These are the angular components of the tracked vehicle configuration; and It is the position component of the target configuration of the tracked vehicle; and It is the position component of the starting configuration of the tracked vehicle.

3. The path planning method for unmanned tracked vehicles based on the improved RRT algorithm of kinematics as described in claim 2, characterized in that: Step S200 includes, S210, Set target bias probability During each sampling, a probability is generated in a uniform distribution between 0 and 1. ,when Generate random configurations of tracked vehicles within the map space, or select target configurations elsewhere. This serves as the configuration for this random sampling. S220, when Generate random configurations within the map space. : （2） In the formula, For map space The range of components, For map space The range of components; and The positional components of the randomly generated tracked vehicle configuration; For random configuration angle direction components; To generate randomly according to a uniform distribution The values ​​between.

4. The path planning method for unmanned tracked vehicles based on the improved RRT algorithm of kinematics as described in claim 3, characterized in that: In step S300, when searching for the nearest configuration on the path tree using a random configuration, the distance metric uses the following distance function: （3） In the formula, This is a distance metric function used to measure distance. and The distance between the two configurations; and for Positional components of the configuration for angular components of the configuration; and for Positional components of the configuration for angular components of the configuration; for and The smaller one.

5. The path planning method for unmanned tracked vehicles based on the improved RRT algorithm of kinematics according to claim 4, characterized in that: In step S400, the following are included: S410. After obtaining the configuration closest to the random configuration on the path tree through step S300, calculate the Euclidean distance between the positional components of the nearest configuration and the destination configuration. ; S420, Euclidean distance between the positional components in the nearest and final configurations obtained through S410. Then, make a judgment Is it greater than the cutoff distance? If it is less than, then a control set scaling factor is given. and If it is greater than 1, then calculate the control set scaling factor: （4） In the formula, It is a control set Scale factor; It is a control set Scale factor; Indicates to Round down to the nearest integer.

6. The path planning method for unmanned tracked vehicles based on the improved RRT algorithm of kinematics according to claim 5, characterized in that: In step S500, the following are included: S510, Provide the control set for tracked vehicles. Obtain its speed control subset and angular velocity control subset and speed control subset The non-negative portion serves as a new velocity control subset. Perform subsequent sampling; S520: The control set scaling factor obtained through S420 is combined with the control set to obtain the final control vector by random sampling according to a uniform distribution. （5） In the formula, The control vector obtained by sampling, Its velocity component, Its angular velocity component; S530, the control vector obtained through S520, is then integrated within a given time according to the tracked vehicle state transition equation to generate a new configuration: （6） In the formula, The new configuration obtained by integration; It is the integration time; and New configuration for tracked vehicles Positional components; For the angular direction components of the new configuration; and For the closest positioning of the tracked vehicle Positional components; The nearest configuration angle direction component; set up The preset threshold parameter is used when the sampled data is obtained. Value at When the interval is within a certain range, the formula for calculating the new configuration is: （7）。 7. The path planning method for unmanned tracked vehicles based on the improved RRT algorithm of kinematics as described in claim 6, characterized in that: In step S700, the target configuration neighborhood is represented as follows: （8） In the formula, The target configuration neighborhood range; The radius of the allowable positional error for the target configuration; The radius of the allowable angular error for the target configuration.

8. A path planning system for unmanned tracked vehicles based on an improved RRT algorithm of kinematics, characterized in that: The system has a program module corresponding to the steps of any one of the claims 1-7 above, and executes the steps in the above-described unmanned tracked vehicle path planning method based on the improved RRT algorithm of kinematics when it is run.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program configured to, when invoked by a processor, implement the steps of the unmanned tracked vehicle path planning method based on the improved RRT algorithm of kinematics as described in any one of claims 1-7.

Images