Path planning method and device, vehicle, storage medium and program product

Abstract

This application relates to a path planning method, apparatus, vehicle, storage medium, and program product. It generates an initial path for a target vehicle based on a target start point and a target end point, and divides the initial path into multiple segments based on the curvature at multiple locations along the initial path. It calculates the attractive and repulsive potential field gradients for each segment, and calculates the potential force gradients corresponding to the multiple segments based on these gradients. Based on the potential force gradients and the multiple segments, it generates multiple planned road segments that meet preset global requirements, thereby generating a planned path for the target vehicle. This application improves the algorithm's convergence speed and computational efficiency by dividing the initial path into multiple segments, allowing each segment to be optimized within a smaller elliptical region. Through a hybrid guidance mechanism, it avoids segments getting trapped in local optima, effectively ensuring the high quality of the generated planned path while improving convergence speed.

Description

Technical Field

[0001] This application relates to the field of path planning technology, and more specifically, to a path planning method, apparatus, vehicle, storage medium, and program product. Background Technology

[0002] Path planning is one of the core technologies of autonomous systems such as mobile robots, autonomous vehicles and drones. Its goal is to find the optimal or feasible path from the starting point to the target point in an environment with obstacles.

[0003] Among related technologies, sampling-based path planning algorithms have attracted widespread attention due to their effectiveness in high-dimensional spaces. They can effectively improve the convergence speed of path optimization by restricting the sampling space to an elliptical region with the starting point and the target point as the focus and the current optimal path length as the major axis.

[0004] However, in related technologies, this path planning method results in its sampling ellipse region covering almost the entire map, losing the advantage of sampling focus and significantly reducing the convergence speed; in narrow passages or areas with many obstacles, it is easy to get stuck in local optima and it is difficult to get out quickly; moreover, this method relies on random sampling and lacks active guidance for the target direction, generating a large number of invalid extended nodes, which greatly reduces the search efficiency and urgently needs improvement. Summary of the Invention

[0005] The purpose of this application is to provide a path planning method, apparatus, vehicle, storage medium, and program product to solve the technical problems in related technologies, such as the path planning method being prone to producing excessively large sampling elliptical regions, resulting in a significant decrease in convergence speed, and being prone to getting stuck in local optima in narrow channels or areas with many obstacles, making it difficult to quickly escape; and the random sampling being blind, lacking active guidance for the target direction, generating a large number of invalid extended nodes, and significantly reducing search efficiency.

[0006] To achieve the above objectives, according to one aspect of this application, a path planning method is provided, comprising the following steps: An initial path for the target vehicle is generated based on its target start point and target end point, and the initial path is divided into multiple road segments based on the curvature of multiple locations on the initial path. Calculate the attractive potential field gradient and the repulsive potential field gradient for each of the plurality of road segments, and calculate the potential force gradient corresponding to the plurality of road segments based on the attractive potential field gradient and the repulsive potential field gradient; Based on the potential field force gradient and the multiple road segments, multiple planned road segments that meet preset global requirements are generated, so as to generate the planned path of the target vehicle based on the multiple planned road segments.

[0007] Optionally, in one embodiment of this application, the step of generating an initial path for the target vehicle based on the target starting point and target ending point, and dividing the initial path into multiple road segments based on the curvature of multiple positions on the initial path, includes: traversing multiple path nodes in the initial path, determining the second node that satisfies a first preset condition and the second-to-last node that satisfies a second preset condition, and calculating the approximate curvature value of the middle node among three consecutive nodes between the second node and the second-to-last node; when the approximate curvature value is greater than or equal to a preset curvature threshold, dividing the initial path using the middle node as a road segmentation point to obtain the multiple road segments.

[0008] Optionally, in one embodiment of this application, the step of dividing the initial path using the intermediate node as a segmentation point to obtain the plurality of road segments includes: calculating the Euclidean distance between the intermediate node and the node preceding the intermediate node; when the Euclidean distance is greater than or equal to a preset length, dividing the initial path using the intermediate node as a segmentation point, and using the road segment between the intermediate node and the node preceding the intermediate node as one of the plurality of road segments.

[0009] Optionally, in one embodiment of this application, calculating the attractive potential field gradient and repulsive potential field gradient of each of the plurality of road segments includes: obtaining a set of prior obstacles in the target area corresponding to the plurality of road segments; calculating the attractive potential field gradient of the plurality of road segments based on the attractive force between the target endpoint and the plurality of road segments; and calculating the repulsive potential field gradient of the plurality of road segments based on the repulsive force between each prior obstacle in the set of prior obstacles and the plurality of road segments.

[0010] Optionally, in one embodiment of this application, after generating the planned path of the target vehicle based on the plurality of planned road segments, the method further includes: obtaining real-time environmental information of the target vehicle; determining current obstacles around the target vehicle based on the real-time environmental information; detecting whether the current obstacle belongs to the obstacle in the prior obstacle set; and if the current obstacle does not belong to the obstacle in the prior obstacle distribution, replanning the planned path until the replanned path satisfies the planning objective.

[0011] Optionally, in one embodiment of this application, the step of replanning the planned path when the current obstacle does not belong to the prior obstacle distribution until the replanned path meets the planning objective includes: obtaining the distance between the current obstacle and each node in each planned road segment; when at least one distance is less than a preset distance, determining the planned road segment corresponding to the node whose distance is less than the preset distance; generating a local planning area based on the starting node position and ending node position of the planned road segment, optimizing the planned road segment within the local planning area to meet preset optimization requirements, and generating a new planned road segment; and replacing the planned road segment with the new planned road segment to complete the replanning of the planned path.

[0012] According to a second aspect of this application, a path planning apparatus is provided, comprising: The segmentation module is used to generate an initial path for the target vehicle based on the target start point and target end point, and to divide the initial path into multiple segments based on the curvature of multiple positions on the initial path. The calculation module is used to calculate the attractive potential field gradient and the repulsive potential field gradient of each of the multiple road segments, and to calculate the potential force gradient corresponding to the multiple road segments based on the attractive potential field gradient and the repulsive potential field gradient. The planning module is used to generate multiple planned road segments that meet preset global requirements based on the potential force gradient and the multiple road segments, so as to generate the planned path of the target vehicle based on the multiple planned road segments.

[0013] Optionally, in one embodiment of this application, the partitioning module includes: a traversal unit, configured to traverse multiple path nodes in the initial path, determine a second node that satisfies a first preset condition and a penultimate node that satisfies a second preset condition among the multiple path nodes, and calculate the approximate curvature value of the middle node among three consecutive nodes between the second node and the penultimate node; and a segmentation unit, configured to segment the initial path using the middle node as a segmentation point when the approximate curvature value is greater than or equal to a preset curvature threshold, thereby obtaining the multiple road segments.

[0014] Optionally, in one embodiment of this application, the segmentation unit includes: a calculation subunit, used to calculate the Euclidean distance between the intermediate node and the node preceding the intermediate node; and a segmentation subunit, used to segment the initial path with the intermediate node as a segmentation point when the Euclidean distance is greater than or equal to a preset length, and to use the segment between the intermediate node and the node preceding the intermediate node as one of the multiple segments.

[0015] Optionally, in one embodiment of this application, the calculation module includes: a first acquisition unit, configured to acquire a set of prior obstacles in the target area corresponding to the plurality of road segments; a first calculation unit, configured to calculate the attractive force potential field gradient corresponding to the plurality of road segments based on the attractive force between the target endpoint and the plurality of road segments; and a second calculation unit, configured to calculate the repulsive force potential field gradient corresponding to the plurality of road segments based on the repulsive force between each prior obstacle in the set of prior obstacles and the plurality of road segments.

[0016] Optionally, in one embodiment of this application, it further includes: an acquisition module, configured to acquire real-time environmental information of the target vehicle after generating the planned path of the target vehicle based on the plurality of planned road segments, and determine the current obstacles around the target vehicle based on the real-time environmental information; and a replanning module, configured to detect whether the current obstacle belongs to the obstacle in the prior obstacle set, and if the current obstacle does not belong to the obstacle in the prior obstacle distribution, replan the planned path until the replanned path satisfies the planning objective.

[0017] Optionally, in one embodiment of this application, the replanning module includes: a second acquisition unit, configured to acquire the distance between the current obstacle and each node in each planned road segment, and determine the planned road segment corresponding to the node whose distance is less than the preset distance when at least one of the distances is less than the preset distance; a generation unit, configured to generate a local planning area based on the starting node position and the ending node position of the planned road segment, so as to optimize the planned road segment in the local planning area to meet the preset optimization requirements and generate a new planned road segment; and a replanning unit, configured to replace the planned road segment with the new planned road segment to complete the replanning of the planned path.

[0018] According to a third aspect of this application, a vehicle is provided, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the processor executing the program to implement the path planning method as described in the above embodiments.

[0019] According to a fourth aspect of this application, a computer-readable storage medium is provided that stores a computer program, which, when executed by a processor, implements the path planning method described above.

[0020] According to a fifth aspect of this application, a computer program product is provided, including a computer program that, when executed, implements the path planning method described above.

[0021] The above technical solution divides the initial path of the target vehicle into multiple segments. Within the local region corresponding to each segment, the potential and force gradients of multiple segments are calculated to generate the planned path of the target vehicle. This achieves the goal of maintaining the asymptotic optimality of the global layer while dividing the initial path into multiple segments, allowing each segment to be optimized within a smaller elliptical region, thus improving the algorithm's convergence speed and computational efficiency. Furthermore, this application also avoids segments getting trapped in local optima by combining a hybrid guidance mechanism that integrates swarm intelligence collaboration with global search capabilities and local target orientation with obstacle avoidance capabilities. This effectively ensures the high quality of the generated planned path while improving convergence speed, forming an efficient, robust, and adaptable path planning method. This solves the technical problems in related technologies, such as path planning methods easily generating excessively large sampling elliptical regions, significantly reducing convergence speed; easily getting trapped in local optima in narrow passages or areas with many obstacles, making it difficult to quickly escape; and the blindness of random sampling, lacking active guidance for the target direction, generating a large number of invalid expansion nodes, and significantly reducing search efficiency.

[0022] Other features and advantages of this application will be described in detail in the following detailed description section. Attached Figure Description

[0023] The accompanying drawings are provided to further illustrate the present application and form part of the specification. They are used together with the following detailed description to explain the present application, but do not constitute a limitation thereof. In the drawings: The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 This is a schematic diagram of the architecture of an intelligent path planning system based on a segmented potential field guided by particle swarm optimization, according to one embodiment of this application. Figure 2 This is a flowchart of a path planning method provided according to an embodiment of this application; Figure 3 This is a flowchart of an intelligent path planning method based on a piecewise potential field guided by particle swarm optimization, according to one embodiment of this application. Figure 4 This is a block diagram of a path planning device provided according to an embodiment of this application; Figure 5 This is a structural schematic diagram of a vehicle provided according to an embodiment of this application.

[0024] Explanation of reference numerals in the attached figures 10 - Path planning device; 100 - Partitioning module; 200 - Calculation module; 300 - Planning module; 501 - Memory; 502 - Processor; 503 - Communication interface. Detailed Implementation

[0025] The specific embodiments of this application will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit this application.

[0026] In this application, unless otherwise stated, directional terms such as "up," "down," "top," "bottom," "front," "rear," "left," and "right" are generally defined in the context of a vehicle's normal driving state (see reference for details). Figure 1 (As shown), it is only for the convenience of describing this application and simplifying the description, and is not intended to indicate or imply that the device or element referred to must have a specific orientation, or a specific orientation construction and operation, and therefore should not be construed as a limitation of this application. "Inner" and "outer" refer to the inner and outer contours of the corresponding components. In addition, the terms "first," "second," etc., used are to distinguish one element from another and do not have any order or importance.

[0027] In the description of this application, it should also be noted that, unless otherwise expressly specified and limited, the terms "set up," "connect," "link," and "install" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a direct connection or an indirect connection through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0028] Before introducing the path planning method provided in the embodiments of this application, the system architecture involved in the embodiments of this application will be introduced first.

[0029] Figure 1 This is a schematic diagram of the architecture of an intelligent path planning system based on piecewise potential field guidance according to an embodiment of this application. Figure 1 As shown, the path planning system in this embodiment mainly consists of four main modules: a perception module, a planning module, a decision-making module, and an execution module.

[0030] The system comprises: a perception module for acquiring environmental information; a planning module for executing the core method proposed in this application; a decision module for generating control commands based on the planning results; and an execution module for driving the actuators and drivers to move according to the control commands.

[0031] Specifically, Figure 2 This is a flowchart of a path planning method provided according to an embodiment of this application.

[0032] like Figure 2 As shown, the path planning method includes the following steps: In step S201, an initial path for the target vehicle is generated based on the target starting point and target ending point, and the initial path is divided into multiple road segments based on the curvature of multiple positions on the initial path.

[0033] It is understandable that the target vehicle here can be understood as the specific vehicle object used for route planning. The target start point and target end point can be understood as the starting point and final destination of the vehicle, respectively.

[0034] In some embodiments, after obtaining the target starting point and target ending point of the vehicle, the initial path of the vehicle can be generated based on in-vehicle navigation or other relevant in-vehicle applications.

[0035] In this embodiment, the initial path may be generated by, but is not limited to, Informed RRT* (InformedRapidly-exploring Random Tree Star). Informed RRT* can effectively improve the convergence speed of the initial path optimization by restricting the sampling space to an elliptical region with the target start and target end points as foci and the current optimal path length as the major axis, thereby ensuring the asymptotic optimality of the global layer.

[0036] Understandably, in environments where the initial path is relatively short but the path is winding (such as mazes or winding passages), the initial path length is much greater than the straight-line distance. The sampling ellipse region of Informed RRT* may become too large, thus losing the advantage of sampling focus and causing a significant decrease in convergence speed.

[0037] In order to effectively reduce the sampling area of ​​Informed RRT*, the embodiments of this application can divide the initial path into multiple segments according to the curvature of multiple positions on the initial path, so that each segment can be optimized within a smaller elliptical area, and finally achieve the effect of greatly improving the convergence speed.

[0038] The embodiments of this application can adaptively divide the global initial path of the target vehicle into several road segments according to the curvature characteristics, so that each road segment can be optimized within a smaller elliptical region, thereby effectively solving the problem of excessively large elliptical regions under high curvature paths and improving the convergence speed of the algorithm.

[0039] Optionally, in one embodiment of this application, the initial path is divided into multiple road segments based on the curvature of multiple locations on the initial path, including: traversing multiple path nodes in the initial path, determining the second node that satisfies a first preset condition and the second-to-last node that satisfies a second preset condition, and calculating the approximate curvature value of the middle node among three consecutive nodes between the second node and the second-to-last node; when the approximate curvature value is greater than or equal to a preset curvature threshold, the initial path is divided using the middle node as a road segmentation point to obtain multiple road segments.

[0040] In actual implementation, when this application divides the initial path into multiple segments based on the curvature at multiple locations of the initial path, it mainly uses curvature characteristics to set segmentation points at locations where curvature changes drastically.

[0041] By analyzing the geometric relationship of three consecutive nodes on the path, the approximate curvature value at the intermediate node can be calculated. When the curvature of a node exceeds a preset threshold and the current accumulated segment length meets the minimum segment length requirement, that node is used as a segmentation point to complete the division of a path segment and start a new path segment until the entire initial path is divided.

[0042] For example, this application may, but is not limited to, traverse all path nodes in the initial path and first determine the second node that satisfies the first preset condition and the second-to-last node that satisfies the second preset condition.

[0043] Here, the first preset condition can be understood as a pre-set condition that the second node must meet. In this embodiment, the second node can be set as the second node counting from the first node in the initial path. The second preset condition can be understood as a pre-set condition that the penultimate node must meet. In this embodiment, it can be set as follows: with the last node in the initial path as the penultimate node, the penultimate node is the node preceding the penultimate node.

[0044] In this context, a node can be understood as a sampling point that has been predetermined when generating the initial path. The first node is the target starting point, and the last node is the target ending point. For example, when generating the path, the target starting point is used as the first node, and then a point is sampled every 0.1m / 0.5m until the target ending point is reached, forming a node sequence. These nodes are discrete coordinate points arranged in sequence that make up the entire path.

[0045] Then, in this embodiment of the application, the approximate curvature value of the middle node (i.e., the middle node is located between the second node and the second to last node, including the second node and the second to last node, avoiding the first node and the last node from being unable to serve as the middle node of three consecutive nodes) can be calculated. If the approximate curvature value of the middle node is greater than or equal to a preset curvature threshold, the middle node can be used as a road segmentation point to divide the initial path. By analogy, multiple road segments can be obtained.

[0046] Here, the preset curvature threshold can be understood as a pre-set curvature critical value. When the approximate curvature value of the intermediate node is greater than this value, the curvature of the intermediate node can be considered to be large, and the initial path can be divided based on this node to avoid the sampling elliptical region being too large due to the high curvature path.

[0047] For example, the approximate curvature value calculation in this application uses the three-point method, which calculates the approximate curvature value for three consecutive nodes in a path. The formula for calculating the approximate curvature value of intermediate nodes can be, but is not limited to, expressed as follows: .

[0048] This application embodiment can analyze the geometric relationship of three consecutive nodes on the path to calculate the approximate curvature value at the intermediate node. When the curvature of a node exceeds a certain threshold and the current accumulated segment length meets the minimum segment length requirement, the node is used as a segmentation point to complete the division of a path segment. This allows the initial path to be divided into multiple segments, enabling each segment to be optimized within a smaller elliptical region. This effectively solves the problem of excessively large elliptical regions under high curvature paths, thereby improving the convergence speed of the algorithm.

[0049] Optionally, in one embodiment of this application, the initial path is divided using intermediate nodes as road segmentation points to obtain multiple road segments, including: calculating the Euclidean distance between an intermediate node and the node preceding the intermediate node; when the Euclidean distance is greater than or equal to a preset length, the initial path is divided using intermediate nodes as road segmentation points, and the road segment between an intermediate node and the node preceding the intermediate node is taken as one of the multiple road segments.

[0050] In other embodiments, considering that curvature is the main factor affecting the sampling elliptical region, classifying excessively small road segments as sampling elliptical regions would also reduce the algorithm's convergence speed. Therefore, embodiments of this application can also calculate the Euclidean distance between an intermediate node and its preceding node. Then, when the approximate curvature value of the intermediate node is greater than or equal to a certain threshold, and the Euclidean distance between the intermediate node and its preceding node is greater than or equal to a preset length, the intermediate node is used as the road segmentation point to segment the initial road segment. The road segment between the intermediate node and its preceding node is one such road segment.

[0051] Here, the preset length can be understood as the minimum segment length that each segment must meet when dividing the initial path. The specific preset length can be set or adjusted by those skilled in the art according to actual needs. This embodiment is only illustrative and does not impose any specific limitations.

[0052] For example, Figure 3 This is a flowchart of an embodiment of the intelligent path planning method based on piecewise potential field guidance of particle swarm optimization, as shown below. Figure 3 As shown, the process of dividing the initial path into multiple road segments in this embodiment of the application can be represented, but is not limited to, as follows: First, we can initialize an empty list of path segments, and use the starting point of the path as the starting point of the current segment, while also initializing the length of the current segment to 0. Then, starting from the second node of the path, traverse sequentially to the second-to-last node, performing the following operations on each intermediate node: Extract the previous node, the current node, and the next node of the current node to form three consecutive nodes, and calculate the approximate curvature value at the current node according to the approximate curvature value calculation formula; Calculate the Euclidean distance from the current node to the previous node and add it to the current segment length; Add the current node to the current segment; Determine whether the curvature at the current node exceeds the preset curvature threshold, and whether the current segment length meets the minimum segment length requirement; If both conditions are met, the current segment is added to the segmentation list as a complete path segment, and a new segment begins with the current node as the new starting point, while the current segment length is reset. If the curvature at the current node does not exceed the preset curvature threshold, or the current segment length does not meet the minimum segment length requirement, no segmentation operation is performed, and the traversal continues to the next node. At this point, the current node has been added to the current segment, and the segment length has been accumulated, so subsequent nodes can continue to be processed sequentially until the segmentation conditions are met or the traversal ends.

[0053] After traversing the path, add the last node of the path to the current segment and add the last segment to the segment list.

[0054] After processing, the initial path can be divided into several path segments. Since the segmentation points are located in areas with drastic curvature changes, the curvature changes within each segment are relatively gentle. This adaptive segmentation method can automatically adjust the number of segments and the length of each segment based on the geometric characteristics of the path, providing a good foundation for subsequent segmentation optimization.

[0055] In step S202, the attractive force potential field gradient and the repulsive force potential field gradient of each of the multiple road segments are calculated, and the potential force gradient of the multiple road segments is calculated based on the attractive force potential field gradient and the repulsive force potential field gradient.

[0056] In some embodiments, after dividing the initial path into multiple segments, in order to improve the search efficiency and path quality, this application may, but is not limited to, combine swarm intelligence search optimized by particle swarm optimization (PSO) with local guidance by artificial potential field (APF) to form a hybrid guidance mechanism, thereby effectively ensuring path quality while improving path search efficiency.

[0057] For example, this application can treat each road segment as a particle, and the multiple road segments corresponding to the initial path form a particle swarm. Each particle represents a road segment and continuously updates its own position (i.e., the position of the path node) by tracking the individual historical best and the group's global best.

[0058] Then, in the embodiment of this application, an artificial potential gradient can be introduced into the particle position update. The artificial potential gradient is calculated based on the current particle position, the target endpoint position and the prior obstacle, so that the particle is attracted by the target endpoint and repelled by the obstacle. Then, the population search of PSO and the local guidance of APF are balanced by adaptive weight η.

[0059] The calculation of the artificial potential field gradient is divided into two parts: the calculation of the attractive gradient and the calculation of the repulsive gradient. The attractive potential field gradient corresponds to the attraction of each road segment to the target endpoint, while the repulsive potential field gradient corresponds to the repulsion of each road segment by the prior obstacles.

[0060] The vector sum of the attractive force gradient and the repulsive force gradient is the artificial potential field force gradient, which is the potential field force gradient in the embodiments of this application.

[0061] Optionally, in one embodiment of this application, calculating the attractive potential field gradient and repulsive potential field gradient of each of the multiple road segments includes: obtaining a set of prior obstacles in the target area corresponding to the multiple road segments; calculating the attractive potential field gradient of the multiple road segments based on the attractive force between the target endpoint and the multiple road segments; and calculating the repulsive potential field gradient of the multiple road segments based on the repulsive force between each prior obstacle in the prior obstacle set and the multiple road segments.

[0062] Based on the descriptions of other embodiments, it can be understood that in the embodiments of this application, the attractive potential field gradient of each road segment corresponds to the attraction of the target endpoint received by each road segment, and the repulsive potential field gradient of each road segment corresponds to the repulsion of the prior obstacle received by each road segment.

[0063] In this context, prior obstacles can be understood as fixed obstacle information obtained in advance through maps, geographic information system (GIS) data, historical experience, or manual annotation. Examples include static obstacles such as buildings, curbs, medians, bollards, fixed construction barriers, and large roadside facilities; impassable areas such as areas outside road boundaries, flower beds, and prohibited areas; and prior risk areas such as accident-prone areas and areas prone to flooding identified from historical data.

[0064] For example, this application can first obtain a set of prior obstacles in the target area corresponding to multiple road segments. Here, the target area can be understood as an elliptical region with the current optimal path length of each road segment as its major axis. The set of prior obstacles can be understood as the set of prior obstacles existing in the target area corresponding to all road segments.

[0065] Based on the set of prior obstacles existing in the target area corresponding to all road segments, the repulsive force between each prior obstacle and each road segment can be calculated. The total repulsive force gradient is obtained by summing the repulsive forces generated by all obstacles, which is the repulsive force potential field gradient of each road segment.

[0066] The attraction between the target endpoint and each road segment can be directly calculated as the gradient of the attraction potential field corresponding to each road segment.

[0067] For example, as Figure 2 As shown, the calculation process of the potential field force gradient in this application can be represented, but is not limited to, as follows: First, the artificial potential field gradient is calculated based on the current particle position (one particle corresponds to one road segment), the target endpoint position, and the prior obstacle distribution. The calculation of the artificial potential field gradient can be divided into two parts: the attractive force gradient calculation and the repulsive force gradient calculation. The attraction gradient calculation is based on the attraction of the target endpoint to the particle. Its direction is from the particle to the target point, and its magnitude is proportional to the distance from the particle to the target point. The attraction gradient corresponding to the attraction of the target endpoint to each particle is the attraction potential field gradient of each road segment in this application embodiment. The repulsive force gradient calculation is based on the repulsive force exerted by obstacles on particles. For each obstacle, when the distance from the particle to the obstacle is less than a preset influence range, a repulsive force is generated from the obstacle towards the particle, the magnitude of which is inversely proportional to the distance from the particle to the obstacle. The total repulsive force gradient obtained by summing the repulsive forces generated by all obstacles on each particle is the repulsive force potential field gradient for each road segment. The vector sum of the attractive and repulsive gradients is the artificial potential field force gradient.

[0068] The embodiments of this application can combine particle swarm optimization and artificial potential fields in path planning. Through the cooperation and complementary advantages of particle swarm optimization and artificial potential field guidance, the global search capability of particle swarm optimization is maintained, while the local obstacle avoidance and target orientation capabilities are enhanced. This ensures that the application can efficiently search for high-quality paths in complex environments.

[0069] In step S203, multiple planned road segments that meet preset global requirements are generated based on the potential field force gradient and multiple road segments, so as to generate the planned path of the target vehicle based on the multiple planned road segments.

[0070] As one possible approach, after calculating the potential field force gradient of each particle (each road segment), the embodiments of this application can generate multiple planned road segments that meet preset global requirements based on the potential field force gradient and multiple road segments. Combining these multiple planned road segments can generate the planned path of the target vehicle.

[0071] Specifically, in this embodiment, the position of each particle can be updated based on the potential field gradient of each particle to obtain the updated particle position. The update of each particle position is equivalent to the update of each road segment, and the updated particle position is equivalent to the new planned road segment.

[0072] The preset global requirement here can be understood as the requirement that the new multiple planned road segments must satisfy the condition that the planned path composed of them is the optimal global path. In the embodiments of this application, the global optimization mechanism inherent in particle swarm optimization can directly and effectively ensure that the multiple planned road segments corresponding to the updates of multiple particles can effectively achieve the optimal global path.

[0073] For example, as Figure 2As shown, after obtaining the artificial potential field gradient determined by the vector sum of the attractive and repulsive potential field gradients, this application can dynamically adjust the weight of the artificial potential field guidance coefficient according to the distance of the particle to the nearest obstacle in the prior obstacle set: when the particle is close to the nearest obstacle, the value of the artificial potential field guidance coefficient is increased, making the particle more inclined to follow the potential field guidance to avoid the obstacle; when the particle is far away from the nearest obstacle, the value of the artificial potential field guidance coefficient is decreased, making the particle rely more on particle swarm optimization for searching. Next, the particle velocity can be updated in this embodiment. It should be noted that "velocity" here does not refer to physical speed, but rather a vector that guides the particle's direction and step size in the search space (i.e., the solution space, which is a set of all possible solutions to the problem). Velocity updating essentially involves a weighted sum of these influencing vectors to obtain a new velocity vector, representing "which direction and what step size the particle should move next." The particle velocity update in this embodiment mainly considers the following four factors: Inertia term: The particle's velocity at the previous moment, multiplied by the inertia weight, reflects the particle's inertia. A larger inertia weight in the early stages can enhance the global exploration capability and prevent particles from gathering in local optimal regions too early; a smaller inertia weight in the later stages can strengthen the local fine search and improve the convergence accuracy. Cognitive term: The difference between the particle's historical best position and its current position, multiplied by the cognitive learning factor and a random number, reflects the particle's tendency to learn towards the best position it finds. That is, in the early stages of the search, the cognitive learning factor is appropriately increased to encourage the particle to explore its individual best position extensively, while the random number can ensure that it can escape the drawbacks of local optima. Social term: The difference between the group's global optimal position and the current position, multiplied by the social learning factor and a random number, reflects the tendency of particles to learn towards the optimal position found by the group; that is, in the later stages of the search, increasing the social learning factor promotes the group to converge to a high-potential region, and the random number can ensure that the group can escape the drawback of local optima. Artificial potential field guidance term: The force gradient of the artificial potential field, multiplied by a dynamically adjusted guidance coefficient, reflects the guiding effect of the potential field on the particle. Preventing particles from getting trapped in local extrema near obstacles: Increasing the guidance weight when approaching an obstacle to force avoidance; decreasing the weight when moving away, allowing the PSO's random search mechanism to dominate, avoiding the "local traps" generated by the potential field from dominating particle behavior.

[0074] The updated speed needs to be subject to boundary constraints to ensure that the speed does not exceed the preset maximum speed, so as to avoid the particles moving too fast and missing the optimal solution.

[0075] After the particle's velocity is updated, the embodiments of this application can update the particle's position based on the updated velocity, that is, add the velocity vector to the current position. Since the particle's position is a coordinate point in the search space (e.g., a node on a path, or a set of parameters), and the velocity represents the displacement vector from the current position, the particle's new position = current position + velocity vector.

[0076] In other words, the particle moves a certain distance along the direction of its velocity from its current position to a new position. This allows the particle to gradually search the search space and approach the optimal solution. If the velocity vector is not added to the current position, the particle will not move, and the search cannot proceed.

[0077] Finally, boundary constraints are applied to the updated particle positions to ensure that the particle positions are within a preset elliptical sampling area, that is, within an elliptical area with the current optimal path length of the road segment corresponding to the particle as the major axis, so as to ensure the effectiveness of the sampling.

[0078] This application embodiment can start with an initial path and gradually optimize the quality of each road segment through multiple iterative iterations, ultimately converging the overall path to the optimal solution. In each iteration, all road segments undergo local optimization in sequence, including updating the elliptical sampling region of each segment, calculating the artificial potential field force of each segment, driving the particle swarm to collaboratively sample within each segment, guiding the growth of the RRT* tree based on the potential field, and updating the optimal information of the particles. This progressive iterative optimization approach enables this application to effectively adapt to various complex driving environments.

[0079] After one iteration, all optimized segments are reassembled into a complete path and its cost is evaluated. If the convergence condition is met, the optimization process is terminated early; otherwise, the next iteration continues. This incremental optimization mechanism ensures that the path quality gradually improves with the number of iterations until the preset convergence criterion (converging to the optimal or near-optimal solution) is reached.

[0080] Therefore, the embodiments of this application can optimize each road segment independently by adaptive segmentation based on path curvature characteristics, reduce the sampling area, and effectively improve the convergence speed and calculation speed of path planning. At the same time, by dynamically balancing the two guidance mechanisms PSO and APF, a hybrid guidance mechanism is formed that combines the collective intelligent collaboration and global search capabilities provided by PSO with the local target guidance and obstacle avoidance capabilities provided by APF. This can effectively ensure the high quality of the generated planned path while improving the convergence speed.

[0081] Optionally, in one embodiment of this application, after generating the planned path of the target vehicle based on multiple planned road segments, the method further includes: obtaining real-time environmental information of the target vehicle, determining the current obstacles around the target vehicle based on the real-time environmental information; detecting whether the current obstacle belongs to the obstacle in the prior obstacle set, and if the current obstacle does not belong to the obstacle in the prior obstacle distribution, replanning the planned path until the replanned path meets the planning objective.

[0082] In other embodiments, considering that the information for the planned path comes from existing prior map information, in order to ensure that the planned path always conforms to the actual driving environment of the target vehicle, this application can obtain the real-time environmental information of the target vehicle after generating the planned path of the target vehicle.

[0083] Based on the real-time environmental information of the target vehicle, the embodiments of this application can clearly identify the real environment around the target vehicle, thereby detecting whether the current obstacle in the environment belongs to the obstacle set in the prior knowledge. If all the current obstacles belong to the obstacle set in the prior knowledge, it proves that the planned path can continue to be used. If a current obstacle that does not belong to the obstacle set in the prior knowledge appears in the environment, the artificial potential field of some road sections may be changed due to the current obstacle.

[0084] At this point, the embodiments of this application can perform local optimization on the planned road segments affected by the current obstacles that do not belong to the prior obstacle set in the real-time environment, that is, replan the already generated planned path until the replanned path meets the planning objective.

[0085] Here, the planning objective can be understood as the requirements that the replanned path should achieve. For example, in the replanned path, the current driving segment is the fastest feasible path found, and the path formed by the remaining segments, excluding the already driven segments and the current driving segment, is the globally optimal path for the current environment.

[0086] Optionally, in one embodiment of this application, when the current obstacle is not an obstacle in the prior obstacle distribution, the planned path is replanned until the replanned path meets the planning objective. This includes: obtaining the distance between the current obstacle and each node in each planned road segment; when there is at least one distance less than a preset distance, determining the planned road segment corresponding to the node with the distance less than the preset distance; generating a local planning area based on the starting and ending node positions of the planned road segment, optimizing the planned road segment within the local planning area to meet preset optimization requirements, and generating a new planned road segment; replacing the planned road segment with the new planned road segment to complete the replanning of the planned path.

[0087] In some embodiments, after determining that there are current obstacles that do not belong to the prior obstacle set based on the real-time environmental information of the target vehicle, this application can first obtain the distance between the current obstacles that do not belong to the prior obstacle set and each node in each of the multiple planned road segments.

[0088] If the distance is less than the preset distance, it proves that the current obstacle, which is not an obstacle in the prior obstacle set, may affect the planned road segment where the node is located. Therefore, it is necessary to generate a local planning interval based on the starting node position and the ending node position of the planned road segment where the node is located, so as to re-optimize the planned road segment in the local planning interval to meet the preset optimization requirements, generate a new planned road segment, and replace the corresponding original planned road segment with the new planned road segment, thus completing the replanning of the planned path.

[0089] Here, the preset distance can be understood as a pre-set minimum distance threshold; road segments smaller than this threshold can be considered affected. In this embodiment, the preset distance can be, but is not limited to, twice the sum of the robot's body radius and the safety margin, taking into account factors such as obstacle size and robot safety distance.

[0090] The preset optimization requirements here can be understood as the best optimization effect that the newly generated planned road segment can achieve, such as the fastest feasible path or a path that can still form the globally optimal path.

[0091] It should be noted that in practical applications, the specific preset distance or preset optimization requirements can be set or adjusted by professionals in this field according to actual needs. The embodiments in this application are only illustrative and do not impose specific limitations.

[0092] For example, the process of replanning the generated planned path based on the real-time environmental information of the target vehicle in this application can be represented, but is not limited to, as follows: First, environmental change detection and affected segment identification are performed. For example, the target vehicle can continuously acquire real-time environmental information through a perception module, including but not limited to data from various sensors such as LiDAR, vision sensors, and ultrasonic sensors. This real-time sensor data is then fused and compared with the original environmental map to detect whether new obstacles have appeared or whether prior obstacles have moved or disappeared. When an environmental change, i.e., a change in an obstacle, is detected, the location coordinates and affected area of ​​the change are recorded.

[0093] Based on environmental change information, the system analyzes which segments of the current optimal path (based on the generated planned path) might be affected. Specifically, for each segment of the current optimal path, the system calculates the shortest distance from all nodes on that segment to the changed location. If any node's distance to the changed location is less than a preset impact threshold (preset distance), then that segment is considered affected.

[0094] If the detection result is empty (i.e. no affected segment exists), it indicates that the current environmental changes do not affect the planned path, the algorithm maintains the current optimal path, and no replanning is required.

[0095] If any road segments are affected, proceed to the local replanning phase. For each affected road segment, perform the following sub-steps: Obtain the start and end node positions of the segment from the path segment data structure, and use them as target point pairs for local replanning; The local planning area extends outward from the affected segment as its center. This extension range can be dynamically determined based on factors such as the segment length and obstacle density, typically ranging from 1.5 to 2 times the segment length, to ensure sufficient adjustment space for the new planned path. In practical applications, the extension range can be set or adjusted by those skilled in the art according to actual needs. This embodiment is merely illustrative and does not impose specific limitations.

[0096] Furthermore, within this expanded local planning area, embodiments of this application can extract all obstacle information from the current obstacle map to form a local obstacle map. Based on this local obstacle map, embodiments of this application can perform rapid path optimization within the local planning area using a simplified version of the path planning method described in this application.

[0097] For example, within this local planning area, the embodiments of this application can reduce the number of particles in particle swarm optimization, for example, by setting it to 30% to 50% of the standard version, thereby reducing the number of iterations to meet the requirements of fast response, simplifying the calculation of artificial potential fields, considering only obstacles within the local area, and prioritizing the rapid finding of feasible paths rather than strictly pursuing the optimal.

[0098] The reduction in the number of particles and the reduction in the number of iterations can be set or adjusted by those skilled in the art according to actual needs. The embodiments in this application are only illustrative and do not impose specific limitations.

[0099] The local path is initialized as a straight line segment from the start point to the end point. Within a preset number of iterations, candidate nodes are continuously generated in the local planning area. The quality of the candidate nodes is then evaluated (considering factors such as path length, obstacle avoidance effect, and smoothness). Finally, the optimal candidate node is selected and inserted into the local path, thus forming the optimized local road segment.

[0100] The optimized local road segments replace the affected road segments in the original planned route, forming an updated complete planned route. After the local replanning is completed, the route smoothing stage begins. The updated complete route is smoothed to eliminate possible broken lines, sharp turns, and discontinuities, improving the route's feasibility and passenger comfort.

[0101] The smoothing process can employ various methods, such as Bézier curve fitting, B-spline interpolation, or gradient descent-based path smoothing algorithms. The specific methods can be set or adjusted by those skilled in the art according to actual needs. The embodiments in this application are merely illustrative and do not impose any specific limitations.

[0102] Finally, the global optimal path is updated, and the smoothed path is used as the new global optimal path. Related data structures (such as path cost, path node list, etc.) are also updated synchronously for subsequent navigation and control.

[0103] This application embodiment utilizes a sensor data fusion mechanism to perceive environmental changes in real time and update the obstacle map, maintaining the accuracy of environmental information and providing reliable basic data for path planning. Then, based on real-time environmental information, it identifies affected path segments, performs local replanning, and dynamically adjusts algorithm parameters to achieve adaptive path planning in complex dynamic environments. This allows the application to quickly adapt to environmental changes, especially when dynamic obstacles appear, enabling rapid replanning and ensuring the real-time performance and safety of path planning.

[0104] Furthermore, the embodiments of this application can dynamically adjust the weight parameters of particle swarm optimization and artificial potential field according to the environmental complexity, enabling the algorithm to adaptively balance exploration and development, and adapt to environments of varying complexity. Environmental complexity can be evaluated based on factors such as obstacle density and distribution uniformity; the higher the complexity, the more the algorithm tends to increase the weight of the artificial potential field to enhance obstacle avoidance capabilities.

[0105] According to the path planning method proposed in this application, the initial path of the target vehicle can be divided into multiple road segments. Within the local region corresponding to each road segment, the potential force gradients of multiple road segments are calculated to generate the planned path of the target vehicle. This achieves the goal of maintaining the asymptotic optimality of the global layer while dividing the initial path into multiple road segments, allowing each segment to be optimized within a smaller elliptical region, thereby improving the algorithm's convergence speed and computational efficiency. Furthermore, this application can also avoid road segments getting trapped in local optima by combining a hybrid guidance mechanism that integrates swarm intelligence collaboration and global search capabilities with local target guidance and obstacle avoidance capabilities. This effectively ensures the high quality of the generated planned path while improving the convergence speed, forming an efficient, robust, and adaptable path planning method. This solves the technical problems in related technologies, such as the path planning method easily generating excessively large sampling elliptical regions, resulting in a significant decrease in convergence speed; the tendency to get trapped in local optima in narrow passages or areas with many obstacles, making it difficult to quickly escape; and the blindness of random sampling, lacking active guidance for the target direction, generating a large number of invalid expansion nodes, and significantly reducing search efficiency.

[0106] Next, the path planning apparatus according to the embodiments of this application is described with reference to the accompanying drawings.

[0107] Figure 4 This is a block diagram of a path planning device provided according to an embodiment of this application.

[0108] like Figure 4 As shown, the path planning device 10 includes: a division module 100, a calculation module 200, and a planning module 300.

[0109] The segmentation module 100 is used to generate the initial path of the target vehicle based on the target starting point and target ending point, and to divide the initial path into multiple segments based on the curvature of multiple positions on the initial path.

[0110] The calculation module 200 is used to calculate the attractive potential field gradient and the repulsive potential field gradient of each of the multiple road segments, and to calculate the potential force gradient of the multiple road segments based on the attractive potential field gradient and the repulsive potential field gradient.

[0111] The planning module 300 is used to generate multiple planned road segments that meet preset global requirements based on the potential force gradient and multiple road segments, so as to generate the planned path of the target vehicle based on the multiple planned road segments.

[0112] Optionally, in one embodiment of this application, the partitioning module 100 includes: a traversal unit and a segmentation unit.

[0113] The traversal unit is used to traverse multiple path nodes in the initial path, determine the second node that satisfies the first preset condition and the second-to-last node that satisfies the second preset condition, and calculate the approximate curvature value of the middle node among the three consecutive nodes between the second node and the second-to-last node.

[0114] The segmentation unit is used to divide the initial path into multiple road segments by using the intermediate node as the segmentation point when the approximate curvature value is greater than or equal to the preset curvature threshold.

[0115] Optionally, in one embodiment of this application, the segmentation unit includes: a calculation subunit and a segmentation subunit.

[0116] The calculation sub-unit is used to calculate the Euclidean distance between the intermediate node and the node preceding the intermediate node.

[0117] The segmentation sub-unit is used to divide the initial path with the intermediate node as the segmentation point when the Euclidean distance is greater than or equal to the preset length, and to take the segment between the intermediate node and the node before the intermediate node as one of the multiple segments.

[0118] Optionally, in one embodiment of this application, the calculation module 200 includes: a first acquisition unit, a first calculation unit, and a second calculation unit.

[0119] The first acquisition unit is used to acquire a set of prior obstacles for the target area corresponding to multiple road segments.

[0120] The first calculation unit is used to calculate the gradient of the attraction potential field corresponding to multiple road segments based on the attraction between the target endpoint and multiple road segments.

[0121] The second calculation unit is used to calculate the gradient of the repulsive force potential field corresponding to multiple road segments based on the repulsive force between each prior obstacle in the prior obstacle set and multiple road segments.

[0122] Optionally, in one embodiment of this application, it further includes: an acquisition module and a replanning module.

[0123] The acquisition module is used to acquire the real-time environmental information of the target vehicle after generating the planned path of the target vehicle based on multiple planned road segments, and to determine the current obstacles around the target vehicle based on the real-time environmental information.

[0124] The replanning module is used to detect whether the current obstacle belongs to the prior obstacle set. If the current obstacle does not belong to the prior obstacle distribution, the replanning module replans the planned path until the replanned path meets the planning objective.

[0125] Optionally, in one embodiment of this application, the replanning module includes: a second acquisition unit, a generation unit, and a replanning unit.

[0126] The second acquisition unit is used to acquire the distance between the current obstacle and each node in each planned road segment, and when there is at least one distance less than a preset distance, it determines the planned road segment corresponding to the node whose distance is less than the preset distance.

[0127] The generation unit is used to generate a local planning area based on the starting and ending node positions of the planned road segment, so as to optimize the planned road segment within the local planning area to meet the preset optimization requirements and generate a new planned road segment.

[0128] A replanning unit is used to replace a planned road segment with a new planned road segment in order to complete the replanning of the planned route.

[0129] It should be noted that the foregoing explanation of the path planning method embodiment also applies to the path planning device of this embodiment, and will not be repeated here.

[0130] According to the path planning apparatus proposed in this application, the initial path of the target vehicle can be divided into multiple road segments. Within the local region corresponding to each road segment, the potential force gradients of multiple road segments are calculated to generate the planned path of the target vehicle. This achieves the goal of maintaining the asymptotic optimality of the global layer while dividing the initial path into multiple road segments, allowing each segment to be optimized within a smaller elliptical region, thereby improving the algorithm's convergence speed and computational efficiency. Furthermore, this application can also avoid road segments getting trapped in local optima by combining a hybrid guidance mechanism that integrates swarm intelligence collaboration with global search capabilities and local target guidance with obstacle avoidance capabilities. This effectively ensures the high quality of the generated planned path while improving the convergence speed, forming an efficient, robust, and adaptable path planning method. This solves the technical problems in related technologies, such as the path planning method easily generating excessively large sampling elliptical regions, resulting in a significant decrease in convergence speed; easily getting trapped in local optima in narrow passages or areas with many obstacles, making it difficult to quickly escape; and the blindness of random sampling, lacking active guidance for the target direction, generating a large number of invalid expansion nodes, and significantly reducing search efficiency.

[0131] Figure 5 This is a schematic diagram of the structure of a vehicle according to an embodiment of this application. The vehicle may include: The memory 501, the processor 502, and the computer program stored on the memory 501 and capable of running on the processor 502.

[0132] When the processor 502 executes the program, it implements the path planning method provided in the above embodiments.

[0133] Furthermore, the vehicle also includes: Communication interface 503 is used for communication between memory 501 and processor 502.

[0134] The memory 501 is used to store computer programs that can run on the processor 502.

[0135] Memory 501 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0136] If the memory 501, processor 502, and communication interface 503 are implemented independently, then the communication interface 503, memory 501, and processor 502 can be interconnected via a bus to complete communication between them. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 5 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0137] Optionally, in a specific implementation, if the memory 501, processor 502, and communication interface 503 are integrated on a single chip, then the memory 501, processor 502, and communication interface 503 can communicate with each other through an internal interface.

[0138] Processor 502 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application.

[0139] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the path planning method described above.

[0140] This application also provides a computer program product, including a computer program that, when executed, implements the path planning method described above.

[0141] The preferred embodiments of this application have been described in detail above with reference to the accompanying drawings. However, this application is not limited to the specific details of the above embodiments. Within the scope of the technical concept of this application, various simple modifications can be made to the technical solution of this application, and these simple modifications all fall within the protection scope of this application.

[0142] It should also be noted that the various specific technical features described in the above embodiments can be combined in any suitable manner without contradiction. In order to avoid unnecessary repetition, this application will not describe the various possible combinations separately.

[0143] Furthermore, various different implementations of this application can be combined in any way, as long as they do not violate the spirit of this application, they should also be regarded as the content disclosed in this application.

Claims

1. A path planning method, characterized in that, Includes the following steps: An initial path for the target vehicle is generated based on its target start point and target end point, and the initial path is divided into multiple road segments based on the curvature of multiple locations on the initial path. Calculate the attractive potential field gradient and the repulsive potential field gradient for each of the plurality of road segments, and calculate the potential force gradient corresponding to the plurality of road segments based on the attractive potential field gradient and the repulsive potential field gradient; Based on the potential field force gradient and the multiple road segments, multiple planned road segments that meet preset global requirements are generated, so as to generate the planned path of the target vehicle based on the multiple planned road segments.

2. The method according to claim 1, characterized in that, The process involves generating an initial path for the target vehicle based on its target start and target end points, and dividing the initial path into multiple segments based on the curvature at multiple locations along the initial path, including: Traverse multiple path nodes in the initial path, determine the second node that satisfies the first preset condition and the second-to-last node that satisfies the second preset condition, and calculate the approximate curvature value of the middle node among the three consecutive nodes between the second node and the second-to-last node. When the approximate curvature value is greater than or equal to a preset curvature threshold, the initial path is divided using the intermediate node as a segmentation point to obtain the multiple road segments.

3. The method according to claim 2, characterized in that, The process of dividing the initial path using the intermediate node as a road segmentation point to obtain the multiple road segments includes: Calculate the Euclidean distance between the intermediate node and the node preceding the intermediate node; When the Euclidean distance is greater than or equal to a preset length, the initial path is divided by the intermediate node as a segmentation point, and the segment between the intermediate node and the node preceding the intermediate node is taken as one of the multiple segments.

4. The method according to claim 1, characterized in that, The calculation of the attractive force potential field gradient and the repulsive force potential field gradient for each of the plurality of road segments includes: Obtain the prior obstacle set for the target area corresponding to the multiple road segments; Based on the attraction between the target endpoint and the multiple road segments, calculate the gradient of the attraction potential field corresponding to the multiple road segments; Based on the repulsive force between each prior obstacle in the prior obstacle set and the multiple road segments, the gradient of the repulsive force potential field corresponding to the multiple road segments is calculated.

5. The method according to claim 4, characterized in that, After generating the planned path for the target vehicle based on the multiple planned road segments, the method further includes: Obtain real-time environmental information of the target vehicle, and determine the current obstacles around the target vehicle based on the real-time environmental information; Detect whether the current obstacle belongs to the prior obstacle set. If the current obstacle does not belong to the prior obstacle distribution, replan the planned path until the replanned path meets the planning objective.

6. The method according to claim 5, characterized in that, When the current obstacle does not belong to the prior obstacle distribution, the planned path is replanned until the replanned path meets the planning objective, including: Obtain the distance between the current obstacle and each node in each planned road segment. When at least one of the distances is less than a preset distance, determine the planned road segment corresponding to the node whose distance is less than the preset distance. Based on the starting and ending node positions of the planned road segment, a local planning area is generated to optimize the planned road segment within the local planning area until it meets the preset optimization requirements, thereby generating a new planned road segment. The planned road segment is replaced with the new planned road segment to complete the replanning of the planned route.

7. A path planning device, characterized in that, include: The segmentation module is used to generate an initial path for the target vehicle based on the target start point and target end point, and to divide the initial path into multiple segments based on the curvature of multiple positions on the initial path. The calculation module is used to calculate the attractive potential field gradient and the repulsive potential field gradient of each of the multiple road segments, and to calculate the potential force gradient corresponding to the multiple road segments based on the attractive potential field gradient and the repulsive potential field gradient. The planning module is used to generate multiple planned road segments that meet preset global requirements based on the potential force gradient and the multiple road segments, so as to generate the planned path of the target vehicle based on the multiple planned road segments.

8. A vehicle, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, the processor executing the program to implement the path planning method as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the path planning method as described in any one of claims 1-6.

10. A computer program product, characterized in that, Includes a computer program, which, when executed, is used to implement the path planning method as described in any one of claims 1-6.

Images