Tunnel path planning method, device and equipment for articulated vehicle and medium

Abstract

Embodiments of the present application provide a path planning method, device, equipment and medium for articulated vehicles in a tunnel. The method comprises: obtaining a tunnel environment point cloud map of a target tunnel and a real-time pose of a target articulated vehicle; determining a tunnel center line and a boundary line according to the tunnel environment point cloud map, and providing a reference for center line guidance. Subsequently, taking the tunnel center line as a guidance reference for path planning, generating a path point sequence set based on the real-time pose, the boundary line, a preset target pose of the target articulated vehicle, and a path planning algorithm. Finally, generating a driving path of the target articulated vehicle in the target tunnel according to the path point sequence set. The method is used to improve the driving safety of articulated vehicles in narrow tunnel spaces.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a path planning method, apparatus, equipment and medium for articulated vehicles in tunnels. Background Technology

[0002] In narrow and confined environments such as mine tunnels, articulated vehicles (such as loaders and trolleys) are widely used in transportation and operation scenarios in such environments due to their flexible articulation steering, minimum turning radius suitable for narrow spaces, and high mobility.

[0003] Currently, for the above scenarios, existing technologies mostly adopt traditional path planning algorithms such as Rapidly-exploring Random Tree (RRT) algorithm, which take the shortest path or obstacle avoidance as the single objective to generate driving trajectories for articulated vehicles.

[0004] However, due to the simple structure and narrow space of tunnels, the driving trajectory generated by traditional path planning algorithms tends to be close to the walls, which makes it difficult to meet the safe driving requirements of articulated vehicles in narrow spaces. Summary of the Invention

[0005] This application provides a path planning method, apparatus, equipment, and medium for articulated vehicles in tunnels, which aims to improve the driving safety of articulated vehicles in narrow tunnel spaces.

[0006] In a first aspect, embodiments of this application provide a path planning method for articulated vehicles in a tunnel, including:

[0007] Acquire the tunnel environment point cloud map of the target tunnel and the real-time pose of the target articulated vehicle.

[0008] Based on the tunnel environment point cloud map, determine the tunnel centerline and boundary line.

[0009] Using the tunnel centerline as the guiding reference for path planning, a set of path point sequences is generated based on real-time pose, boundary line, preset target pose of the target articulated vehicle, and path planning algorithm.

[0010] Based on the path point sequence set, generate the driving path of the target articulated vehicle in the target tunnel; the driving path is the path close to the center line of the tunnel in the spatial distribution of the target tunnel.

[0011] In one possible implementation, in conjunction with the first aspect, using the tunnel centerline as a guiding reference for path planning, and based on the real-time pose, boundary line, preset target pose of the target articulated vehicle, and path planning algorithm, a path point sequence set is generated, including:

[0012] Based on the inherent parameters of the target articulated vehicle, a kinematic model of the articulated vehicle is constructed.

[0013] Based on the tunnel environment point cloud map, real-time pose, and preset target pose, a motion state space is generated.

[0014] Construct an initial path tree using the real-time pose as the root node.

[0015] Using the tunnel centerline as the guiding reference for path planning, the initial path tree is iteratively expanded repeatedly based on the articulated vehicle kinematic model, motion state space, boundary lines, and path planning algorithm until the preset termination conditions are met, thus obtaining the target path tree.

[0016] Based on each node of the target path tree, generate a set of path point sequences.

[0017] In one possible implementation, in conjunction with the first aspect, using the tunnel centerline as the guiding reference for path planning, and based on the articulated vehicle's kinematic model, motion state space, boundary lines, and path planning algorithm, the initial path tree is iteratively expanded repeatedly until a preset termination condition is met to obtain the target path tree, including:

[0018] Randomly sample the motion state space to generate direction reference points.

[0019] Based on the tunnel centerline and boundary line, the comprehensive cost of the path from each existing node in the initial path tree to the direction reference point is calculated using a path planning algorithm, and an extension starting point is selected from each existing node. The comprehensive cost represents the degree to which the path deviates from the tunnel centerline, the degree to which the path is close to the boundary line, and the length of the path.

[0020] Based on the articulated vehicle's kinematic model, direction reference point, tunnel centerline, boundary line, and extension starting point, a new node is generated; the extension starting point is the initial parent node of the new node.

[0021] Based on the tunnel centerline and boundary line, the comprehensive cost of the path from each existing node to the new node is calculated using a path planning algorithm, and the target parent node of the new node is determined from the existing nodes.

[0022] Based on the new node and the target parent node, update the initial path tree; and repeatedly perform the process of randomly sampling the motion state space to generate direction reference points until the preset termination condition is met to obtain the target path tree.

[0023] In one possible implementation, in conjunction with the first aspect, new nodes are generated based on the articulated vehicle kinematic model, orientation reference point, tunnel centerline, boundary line, and extension starting point, including:

[0024] The dynamic expansion step size is obtained based on the direction reference point, the tunnel centerline, and the boundary line.

[0025] Based on the dynamic expansion step size and the articulated vehicle kinematic model, the expansion starting point is expanded along the direction of the direction reference point to generate candidate nodes; the articulated vehicle kinematic model is used to ensure that the expansion process satisfies the kinematic constraints of the target articulated vehicle.

[0026] Based on the boundary line and the kinematic model of the articulated vehicle, the legality of candidate nodes is verified; and the process of dynamically expanding the step size by taking the direction reference point, the tunnel centerline and the boundary line is repeated until the candidate node passes the legality verification.

[0027] Candidate nodes that pass the validity check will be identified as new nodes.

[0028] In one possible implementation, in conjunction with the first aspect, the dynamic expansion step size is obtained based on the direction reference point, the tunnel centerline, and the boundary line, including:

[0029] The boundary distance is obtained based on the starting point and boundary line of the extension.

[0030] The curvature of the centerline is obtained based on the starting point of the extension and the centerline of the tunnel.

[0031] The dynamic expansion step size is obtained based on the preset step size threshold, preset curvature influence factor, boundary distance, and centerline curvature. The dynamic expansion step size is positively correlated with the boundary distance and negatively correlated with the centerline curvature.

[0032] In one possible implementation, in conjunction with the first aspect, generating the travel path of the target articulated vehicle in the target tunnel based on the path point sequence set includes:

[0033] The path point sequence set is segmented to obtain multiple path point segment sets and common connection points between each path point segment set.

[0034] Based on the preset path continuity constraints, common connection points, and segmentation sets of each path point, multiple segmented smooth curves are generated.

[0035] The driving trajectory is obtained by connecting the various segmented smooth curves into a continuous whole.

[0036] Secondly, embodiments of this application provide a path planning device for articulated vehicles in tunnels, comprising:

[0037] The acquisition module is used to acquire the tunnel environment point cloud map of the target tunnel and the real-time pose of the target articulated vehicle.

[0038] The determination module is used to determine the tunnel centerline and boundary lines based on the tunnel environment point cloud map.

[0039] The path planning module is used to generate a set of path point sequences based on the tunnel centerline as the guiding reference, the real-time pose, the boundary line, the preset target pose of the target articulated vehicle, and the path planning algorithm.

[0040] The generation module is used to generate the driving path of the target articulated vehicle in the target tunnel based on the path point sequence set; the driving path is the path close to the center line of the tunnel in the spatial distribution of the target tunnel.

[0041] In one possible implementation, in conjunction with the second aspect, the path planning module is specifically used for:

[0042] Based on the inherent parameters of the target articulated vehicle, a kinematic model of the articulated vehicle is constructed.

[0043] Based on the tunnel environment point cloud map, real-time pose, and preset target pose, a motion state space is generated.

[0044] Construct an initial path tree using the real-time pose as the root node.

[0045] Using the tunnel centerline as the guiding reference for path planning, the initial path tree is iteratively expanded repeatedly based on the articulated vehicle kinematic model, motion state space, boundary lines, and path planning algorithm until the preset termination conditions are met, thus obtaining the target path tree.

[0046] Based on each node of the target path tree, generate a set of path point sequences.

[0047] In one possible implementation, in conjunction with the second aspect, the path planning module is specifically used for:

[0048] Based on the inherent parameters of the target articulated vehicle, a kinematic model of the articulated vehicle is constructed.

[0049] Based on the tunnel environment point cloud map, real-time pose, and preset target pose, a motion state space is generated.

[0050] Construct an initial path tree using the real-time pose as the root node.

[0051] Randomly sample the motion state space to generate direction reference points.

[0052] Based on the tunnel centerline and boundary line, the comprehensive cost of the path from each existing node in the initial path tree to the direction reference point is calculated using a path planning algorithm, and an extension starting point is selected from each existing node. The comprehensive cost represents the degree to which the path deviates from the tunnel centerline, the degree to which the path is close to the boundary line, and the length of the path.

[0053] Based on the articulated vehicle's kinematic model, direction reference point, tunnel centerline, boundary line, and extension starting point, a new node is generated; the extension starting point is the initial parent node of the new node.

[0054] Based on the tunnel centerline and boundary line, the comprehensive cost of the path from each existing node to the new node is calculated using a path planning algorithm, and the target parent node of the new node is determined from the existing nodes.

[0055] Based on the new node and the target parent node, update the initial path tree; and repeatedly perform the process of randomly sampling the motion state space to generate direction reference points until the preset termination condition is met to obtain the target path tree.

[0056] Based on each node of the target path tree, generate a set of path point sequences.

[0057] In one possible implementation, in conjunction with the second aspect, the path planning module is specifically used for:

[0058] Based on the inherent parameters of the target articulated vehicle, a kinematic model of the articulated vehicle is constructed.

[0059] Based on the tunnel environment point cloud map, real-time pose, and preset target pose, a motion state space is generated.

[0060] Construct an initial path tree using the real-time pose as the root node.

[0061] Randomly sample the motion state space to generate direction reference points.

[0062] Based on the tunnel centerline and boundary line, the comprehensive cost of the path from each existing node in the initial path tree to the direction reference point is calculated using a path planning algorithm, and an extension starting point is selected from each existing node. The comprehensive cost represents the degree to which the path deviates from the tunnel centerline, the degree to which the path is close to the boundary line, and the length of the path. The extension starting point is the initial parent node of the new node.

[0063] Based on the dynamic expansion step size and the articulated vehicle kinematic model, the expansion starting point is expanded along the direction of the direction reference point to generate candidate nodes; the articulated vehicle kinematic model is used to ensure that the expansion process satisfies the kinematic constraints of the target articulated vehicle.

[0064] Based on the boundary line and the kinematic model of the articulated vehicle, the legality of candidate nodes is verified; and the process of dynamically expanding the step size by taking the direction reference point, the tunnel centerline and the boundary line is repeated until the candidate node passes the legality verification.

[0065] Candidate nodes that pass the validity check will be identified as new nodes.

[0066] Based on the tunnel centerline and boundary line, the comprehensive cost of the path from each existing node to the new node is calculated using a path planning algorithm, and the target parent node of the new node is determined from the existing nodes.

[0067] Based on the new node and the target parent node, update the initial path tree; and repeatedly perform the process of randomly sampling the motion state space to generate direction reference points until the preset termination condition is met to obtain the target path tree.

[0068] Based on each node of the target path tree, generate a set of path point sequences.

[0069] In one possible implementation, in conjunction with the second aspect, the path planning module is specifically used for:

[0070] The boundary distance is obtained based on the starting point and boundary line of the extension.

[0071] The curvature of the centerline is obtained based on the starting point of the extension and the centerline of the tunnel.

[0072] The dynamic expansion step size is obtained based on the preset step size threshold, preset curvature influence factor, boundary distance, and centerline curvature. The dynamic expansion step size is positively correlated with the boundary distance and negatively correlated with the centerline curvature.

[0073] In one possible implementation, in conjunction with the second aspect, the generation module is specifically used for:

[0074] The path point sequence set is segmented to obtain multiple path point segment sets and common connection points between each path point segment set.

[0075] Based on the preset path continuity constraints, common connection points, and segmentation sets of each path point, multiple segmented smooth curves are generated.

[0076] The driving trajectory is obtained by connecting the various segmented smooth curves into a continuous whole.

[0077] Thirdly, embodiments of this application provide an electronic device, including: a processor, and a memory communicatively connected to the processor.

[0078] The memory stores the instructions that the computer executes.

[0079] The processor executes computer execution instructions stored in memory, causing the processor to perform the first aspect and / or various possible implementations of the first aspect as described above.

[0080] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the first aspect and / or various possible implementations of the first aspect.

[0081] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the first aspect and / or various possible implementations of the first aspect.

[0082] This application provides a path planning method, apparatus, device, and medium for articulated vehicles in tunnels. By acquiring a tunnel environment point cloud map of the target tunnel and the real-time pose of the target articulated vehicle, the tunnel centerline and boundary lines are determined based on the tunnel environment point cloud map, providing a reference for centerline guidance. Subsequently, using the tunnel centerline as the guiding reference for path planning, the path planning algorithm prioritizes nodes close to the centerline. Based on the real-time pose, boundary lines, the preset target pose of the target articulated vehicle, and the path planning algorithm, a path point sequence set is generated. Finally, based on the path point sequence set, the driving path of the target articulated vehicle in the target tunnel is generated. These technical means, through the dynamic correlation between tunnel centerline guidance and algorithm parameters, ensure that the generated driving path closely follows the centerline, thereby improving the driving safety of articulated vehicles in narrow tunnel spaces. Attached Figure Description

[0083] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0084] Figure 1 A schematic diagram illustrating a path planning method for articulated vehicles in a tunnel, as provided in this application.

[0085] Figure 2 A flowchart illustrating a path planning method for articulated vehicles in tunnels provided in this application. Figure 1 ;

[0086] Figure 3 A schematic diagram of the tunnel environment map including the tunnel centerline provided in this application;

[0087] Figure 4 A flowchart illustrating a path planning method for articulated vehicles in tunnels provided in this application. Figure 2 ;

[0088] Figure 5 A schematic diagram of the kinematic geometry of an articulated vehicle's kinematic model;

[0089] Figure 6 A flowchart illustrating a path planning method for articulated vehicles in tunnels provided in this application. Figure 3 ;

[0090] Figure 7 A specific example diagram illustrating a path planning method for articulated vehicles in a tunnel provided in this application;

[0091] Figure 8 A schematic diagram of the structure of a path planning device for an articulated vehicle in a tunnel provided in this application;

[0092] Figure 9 A schematic diagram of the structure of the electronic device provided in this application.

[0093] The accompanying drawings have illustrated specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to specific embodiments. Detailed Implementation

[0094] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0095] The application background of the embodiments of this application will be explained below:

[0096] In confined environments such as mining tunnels, articulated vehicles (such as loaders and trolleys) are widely used for transportation and operations in these environments due to their flexible articulation, minimal turning radius adaptable to narrow spaces, and high maneuverability. Currently, for these scenarios, existing technologies mostly employ traditional path planning algorithms such as the RRT algorithm, using the shortest path or obstacle avoidance as the single objective to generate driving trajectories for articulated vehicles. However, because tunnels have a simple structure and are long and narrow, the driving trajectories generated by traditional path planning algorithms tend to hug the walls, making it difficult to meet the safe driving requirements of articulated vehicles in confined spaces.

[0097] To address the aforementioned problems, the inventors investigated whether a path planning method for articulated vehicles in narrow, confined environments could be constructed by integrating tunnel environment geometric feature guidance with an adaptive search strategy. The inventors proposed a path planning method for articulated vehicles in tunnels. This method involves acquiring a tunnel environment point cloud map of the target tunnel and the real-time pose of the target articulated vehicle. Based on the tunnel environment point cloud map, the tunnel centerline and boundary lines are determined, providing a reference for centerline guidance. Subsequently, using the tunnel centerline as the guiding reference for path planning, the path planning algorithm prioritizes nodes close to the centerline. Based on the real-time pose, boundary lines, the preset target pose of the target articulated vehicle, and the path planning algorithm, a path point sequence set is generated. Finally, based on the path point sequence set, the driving path of the target articulated vehicle in the target tunnel is generated. These techniques, through the dynamic correlation between tunnel centerline guidance and algorithm parameters, ensure that the generated driving path closely follows the centerline, thereby improving the driving safety of articulated vehicles in narrow tunnel spaces.

[0098] Taking the path planning scenario of articulated vehicles in a narrow tunnel environment as an example, combined with Figure 1 This illustrates a specific application scenario of the path planning method for articulated vehicles in tunnels provided in this application. For example... Figure 1 As shown, the specific application scenario of this application includes the rear body 101 and front body 102 of an articulated vehicle, a computing device 103, a lidar 104, a tunnel wall 105, a driving area 106, and a non-driving area 107. The rear body 101 and front body 102 of the articulated vehicle are connected by an articulation plate. The lidar 104 is mounted on the front body 102, and the computing device 103 is mounted on the rear body 101. The tunnel wall 105 divides the area in the figure into a driving area 106 and a non-driving area 107. The articulated vehicle travels in the driving area 106, obtains a three-dimensional spatial point set of the driving area 106 through the lidar 104, and obtains the real-time pose of the vehicle through multimodal sensors on the vehicle. The three-dimensional spatial point set and the real-time pose are then transmitted to the computing device 103. The articulated vehicle generates a driving path in the driving area 106 based on a set of three-dimensional spatial points, real-time pose, and preset target pose using a computing device 103, and achieves automatic driving in a narrow tunnel environment based on this driving path. The driving area 106 is the internal space of the tunnel enclosed by the tunnel wall 105.

[0099] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0100] Figure 2 A flowchart illustrating a path planning method for articulated vehicles in tunnels provided in this application. Figure 1 ,like Figure 2 As shown, the method includes:

[0101] S201. Obtain the tunnel environment point cloud map of the target tunnel and the real-time pose of the target articulated vehicle.

[0102] In this step, a three-dimensional spatial point set of the target tunnel is generated by scanning with a LiDAR scanner on the target articulated vehicle, and a point cloud map of the tunnel environment is constructed based on the three-dimensional spatial point set. Simultaneously, the real-time pose of the target articulated vehicle is acquired using multimodal sensors on the target articulated vehicle.

[0103] S202. Determine the tunnel centerline and boundary line based on the tunnel environment point cloud map.

[0104] In this step, the center line and boundary line of the target tunnel are obtained by extracting features from the point cloud map of the tunnel environment.

[0105] Specifically, a skeleton extraction algorithm is used to extract the tunnel centerline and boundary lines from the tunnel environment point cloud map.

[0106] In one possible implementation, a tunnel environment map is generated based on a tunnel environment point cloud map using a skeleton extraction algorithm. This tunnel environment map has already extracted the tunnel centerline and boundary lines. Figure 3 As shown, the tunnel environment map includes a boundary line 301 and a tunnel centerline 302. The boundary line 301 defines the physical outline of the tunnel and is the core basis for dividing the tunnel's driving area from the non-driving area. The tunnel centerline 302 serves as the geometric center reference of the tunnel and can be used to guide articulated vehicles to travel along the tunnel center, avoiding the problem of the path hugging the wall.

[0107] S203. Using the tunnel centerline as the guiding reference for path planning, and based on the real-time pose, boundary line, preset target pose of the target articulated vehicle, and path planning algorithm, generate a set of path point sequences.

[0108] In this step, a kinematic model of the articulated vehicle is first constructed based on its inherent parameters. A motion state space is then generated based on the tunnel environment point cloud map, real-time pose, and a preset target pose. An initial path tree is constructed using the real-time pose as the root node. Subsequently, using the tunnel centerline as the guiding reference for path planning, the initial path tree is iteratively expanded repeatedly based on the articulated vehicle kinematic model, motion state space, boundary lines, and path planning algorithm until a preset termination condition is met, resulting in the target path tree. Finally, a set of path point sequences is generated based on each node of the target path tree. The preset termination condition can be that an existing node in the initial path tree enters the target region corresponding to the preset target pose.

[0109] S204. Based on the path point sequence set, generate the driving path of the target articulated vehicle in the target tunnel.

[0110] The driving path is the path that is close to the center line of the target tunnel in the spatial distribution of the tunnel.

[0111] In this step, the path point sequence set is first segmented to obtain multiple path point segment sets and common connection points between these segments. Then, based on preset path continuity constraints, common connection points, and each path point segment set, multiple segmented smooth curves are generated. Finally, by connecting these segmented smooth curves into a continuous whole, the driving trajectory is obtained. The preset path continuity constraints can be fifth-order polynomial splines or spiral curves. This fundamentally eliminates curvature abrupt changes that cause vehicle control instability, ensuring the curvature continuity of the final generated driving path. This allows the articulated vehicle to perform path tracking with a smoother and more stable posture, improving control accuracy, ride comfort, and path tracking performance.

[0112] In one possible implementation, the path point sequence set As input, a fifth-order polynomial spline is used to fit the path. First, piecewise fitting is performed, at each... and Generate a quintic polynomial curve between the points; then apply constraints to the piecewise curve at each connection point. This process mandates that the positions, first derivatives (directions), and second derivatives (curvature) of the left and right curve segments be exactly equal. This ensures that the generated path conforms to second-order geometrical continuity (G2 continuity), without any cusps or abrupt curvature changes. Finally, the driving trajectory is generated, and by solving this constrained optimization problem, a final driving path that fully satisfies the kinematic feasibility of the articulated vehicle, exhibits curvature continuity, and is highly smooth is produced. ,in Given the arc length, this trajectory can be directly used as the input to the vehicle's underlying tracking controller.

[0113] This application provides a path planning method for articulated vehicles in tunnels. By acquiring a tunnel environment point cloud map of the target tunnel and the real-time pose of the target articulated vehicle, the tunnel centerline and boundary lines are determined based on the tunnel environment point cloud map, providing a reference for centerline guidance. Subsequently, using the tunnel centerline as the guiding reference for path planning, the path planning algorithm prioritizes nodes close to the centerline. Based on the real-time pose, boundary lines, the preset target pose of the target articulated vehicle, and the path planning algorithm, a path point sequence set is generated. Finally, based on the path point sequence set, the driving path of the target articulated vehicle in the target tunnel is generated. These technical means, through the dynamic correlation between tunnel centerline guidance and algorithm parameters, ensure that the generated driving path closely follows the centerline, thereby improving the driving safety of articulated vehicles in narrow tunnel spaces.

[0114] Figure 4 A flowchart illustrating a path planning method for articulated vehicles in tunnels provided in this application. Figure 2 ,like Figure 4 As shown, in this embodiment... Figure 2 Based on the embodiments, S203 will be described in detail, and this step includes:

[0115] S401. Construct the kinematic model, motion state space, and initial path tree of the articulated vehicle.

[0116] In this step, a kinematic model of the articulated vehicle is constructed based on the inherent parameters of the target articulated vehicle; a motion state space is generated based on the tunnel environment point cloud map, real-time pose, and preset target pose; and an initial path tree is constructed with the real-time pose as the root node.

[0117] In one possible implementation, based on the physical articulated structure of an articulated vehicle, and combined with the circular arc motion assumption during node expansion using a path planning algorithm, a kinematic model of the articulated vehicle is constructed through geometric derivation and kinematic constraint definition. The circular arc motion assumption refers to the local expansion path of the articulated vehicle from the expansion starting point to the candidate node being a circular arc of a fixed radius. The kinematic equations of this articulated vehicle kinematic model are as follows:

[0118]

[0119] in, The equivalent motion length is the equivalent radius of rotation for the articulated vehicle's circular motion. To extend the radius of the arc; This refers to the length of the front of the vehicle. This refers to the length of the rear vehicle body; The sum of steering angles represents the maximum steering amplitude in a single expansion. The heading angle of the front vehicle body; The rear vehicle's heading angle; The hinge angle; For the first The heading angle of the vehicle body; For the first The heading angle of the vehicle body; The heading deviation angle represents the deviation between the extended direction and the current heading.

[0120] Figure 5 This is a schematic diagram of the kinematic geometry of an articulated vehicle's kinematic model, representing the kinematic geometry of the articulated vehicle during the node expansion process in the path planning algorithm. For example... Figure 5 As shown, A is the center point of the rear body of the articulated vehicle, with wheels at the bottom; C is the center point of the front body of the articulated vehicle, with wheels at the bottom; the length of the front body... The distance from the center point C of the front vehicle body to the hinge point B, and the length of the rear vehicle body. The distance from hinge point B to the center point A of the rear vehicle body is given. The front and rear vehicle bodies are connected through hinge point B and can only rotate relative to each other around this point. Two black dashed arcs with center O and radius R represent the assumed local motion trajectories during path tree expansion (the center C of the front vehicle body moves along this arc, and R is not less than the vehicle's minimum turning radius). Point N is the expanded node generated in this expansion (called a candidate node in some embodiments), falling on the arc trajectory. The diagram uses "fixed-radius arc motion" as the core assumption, linking the rigid structure and steering angle of the articulated vehicle with the expanded nodes of the path tree. This provides the geometric basis for deriving the kinematic equations and constraining the feasibility of path expansion.

[0121] S402. Randomly sample the motion state space to generate direction reference points.

[0122] In this step, a uniform random sampling strategy is used to generate sampling points randomly and without bias within the motion state space. These sampling points are then used as direction reference points. These direction reference points provide clear directional guidance for subsequent steps.

[0123] In one possible implementation, a target bias sampling mechanism is introduced to directly sample the target pose with a preset bias probability (e.g., 10%~20%), guide the sampling points to tilt toward the target area, and generate a direction reference point.

[0124] S403. Based on the tunnel centerline and boundary line, calculate the comprehensive cost of the path from each existing node in the initial path tree to the direction reference point through the path planning algorithm, and select the extension starting point from each existing node.

[0125] Among them, the comprehensive cost characterizes the degree to which the path deviates from the tunnel centerline, the degree to which the path approaches the boundary line, and the length of the path.

[0126] In this step, all existing nodes stored in the initial path tree are traversed. Based on the tunnel centerline and boundary line, the comprehensive cost from each existing node in the initial path tree to the current direction reference point is calculated one by one using a path planning algorithm. By comparing the comprehensive cost values ​​of all existing nodes in the initial path tree laterally, the node with the optimal comprehensive cost is selected and determined as the starting point for this iteration expansion.

[0127] S404. Based on the articulated vehicle kinematic model, direction reference point, tunnel centerline, boundary line, and extension starting point, generate new nodes.

[0128] The starting point for expansion is the initial parent node of the new node.

[0129] The articulated vehicle kinematic model is used to ensure that the extended process satisfies the kinematic constraints of the target articulated vehicle.

[0130] In this step, a dynamic expansion step size is first obtained based on the direction reference point, the tunnel centerline, and the boundary line. Then, based on the dynamic expansion step size and the articulated vehicle's kinematic model, the expansion starting point is expanded along the direction reference point to generate candidate nodes. Subsequently, the candidate nodes are validated for legality based on the boundary line and the articulated vehicle's kinematic model; this process of obtaining the dynamic expansion step size based on the direction reference point, tunnel centerline, and boundary line is repeated until a candidate node passes the validity check. Finally, the candidate nodes that pass the validity check are determined as new nodes.

[0131] S405. Based on the tunnel centerline and boundary line, calculate the comprehensive cost of the path from each existing node to the new node using a path planning algorithm, and determine the target parent node of the new node from each existing node.

[0132] In S403 and S405, based on the tunnel centerline and boundary line, the starting point and target parent node are selected by calculating the comprehensive cost. The resulting path will actively move away from the tunnel wall, reserving a larger safety redundancy for articulated vehicles (especially their rear end which is prone to "fishtailing"). This greatly reduces the risk of collisions in narrow environments and improves driving safety.

[0133] In one possible implementation, the node is determined by evaluating the comprehensive cost of the path between the two nodes using a path planning algorithm. The advantages and disadvantages of [these] can be expressed as follows:

[0134]

[0135] in, From the starting point to the node The path length cost. Specifically, when calculating the comprehensive cost of the path from each existing node in the initial path tree to the direction reference point using a path planning algorithm, the aforementioned starting point is an existing node, and the aforementioned nodes... The above-mentioned starting point is the reference point for direction; when calculating the comprehensive cost of the path from each existing node to the new node using the path planning algorithm, the above-mentioned starting point is the existing node, and the above-mentioned node n is the new node.

[0136] The cost of guiding the tunnel centerline is used to penalize nodes that deviate from the tunnel centerline. Its calculation method is nodes To the nearest tunnel centerline point vertical distance The square of, that is By minimizing this cost, the path planning algorithm will proactively explore and select a path closer to the center of the tunnel.

[0137] The obstacle rejection cost is related to the node. Distance to the nearest tunnel wall or obstacle Inversely proportional, It is used to enhance obstacle avoidance.

[0138] , , These are the weighting coefficients for various costs, which can be dynamically adjusted according to safety requirements and tunnel width.

[0139] S406. Update the initial path tree based on the new node and the target parent node.

[0140] In this step, based on the new node and the target parent node, an update operation is performed on the initial path tree: the new node is formally incorporated into the path tree structure, and its parent-child relationship with the target parent node is bound.

[0141] S407. Determine whether the preset termination condition has been met; if yes, execute S408; if no, execute S402~S407.

[0142] S408. Obtain the target path tree, and generate a path point sequence set based on each node of the target path tree.

[0143] In this step, when the path tree iterative expansion meets the preset termination condition, the target path tree is obtained. Based on the parent-child relationship topology stored in advance for each node in the target path tree, the end node falling into the target pose's adjacent region is used as the backtracking starting point. The parent nodes at each level are traversed in reverse order until the root node of the path tree, i.e., the starting node, is reached. All valid nodes obtained from the backtracking are reordered and integrated according to the vehicle's driving direction to generate a complete set of path point sequences connecting the starting point to the target pose.

[0144] This application provides a path planning method for articulated vehicles in tunnels. First, a kinematic model of the articulated vehicle, a motion state space, and an initial path tree are constructed. Then, the motion state space is randomly sampled to generate direction reference points. Based on the tunnel centerline and boundary lines, a path planning algorithm calculates the comprehensive cost of the path from each existing node in the initial path tree to the direction reference points, and selects an extension starting point from each existing node. Subsequently, based on the articulated vehicle kinematic model, direction reference points, tunnel centerline, boundary lines, and extension starting points, new nodes are generated. The kinematic model ensures that the new nodes conform to the actual motion capabilities of the articulated vehicle. Then, based on the tunnel centerline and boundary lines, a path planning algorithm calculates the comprehensive cost of the path from each existing node to the new node, determines the target parent node of the new node from each existing node, and generates the optimal path by calculating the comprehensive cost.

[0145] Subsequently, based on the new node and the target parent node, the initial path tree is updated, and the process of randomly sampling the motion state space and generating direction reference points is repeated until a preset termination condition is met, resulting in the target path tree. Finally, based on each node of the target path tree, a path point sequence set is generated, thereby improving the driving safety of articulated vehicles in narrow tunnel spaces.

[0146] Figure 6 A flowchart illustrating a path planning method for articulated vehicles in tunnels provided in this application. Figure 3 ,like Figure 6 As shown, this embodiment provides a detailed description of S404 based on any of the above embodiments. This step includes:

[0147] S601. Obtain the boundary distance based on the starting point of the extension and the boundary line.

[0148] In this step, the Euclidean distances from the expansion starting point to the left and right boundary lines are calculated, and the smaller value is taken as the boundary distance. This boundary distance directly represents the lateral available space at the location of the expansion starting point and is a key constraint parameter for subsequent dynamic adjustment of the expansion step size. It is used to avoid collisions between the vehicle and the tunnel wall due to excessive step size, and at the same time provides a quantitative basis for the lateral safety margin of the articulated vehicle.

[0149] S602. Based on the starting point of the extension and the centerline of the tunnel, the curvature of the centerline is obtained.

[0150] In this step, the extended starting point is first projected onto the tunnel centerline to determine the corresponding centerline projection point. Then, based on the projection point and its adjacent local centerline point sequence, the centerline curvature at that projection point is calculated using quadratic curve fitting or the finite difference method. This centerline curvature reflects the sharpness of the tunnel curve at the current location: a larger centerline curvature indicates a sharper curve, requiring more adjustment space for vehicle steering. Fine-grained planning through step size adjustments is necessary to avoid subsequently generated nodes exceeding the articulated vehicle's kinematic constraints.

[0151] S603. Based on the preset step size threshold, preset curvature influence factor, boundary distance, and centerline curvature, obtain the dynamic expansion step size.

[0152] Among them, the dynamic expansion step size is positively correlated with the boundary distance and negatively correlated with the centerline curvature. Specifically, the positive correlation between the dynamic expansion step size and the boundary distance means that when the expansion starting point is far from the tunnel wall, the dynamic expansion step size increases to improve search efficiency; the negative correlation between the dynamic expansion step size and the centerline curvature means that when the expansion starting point is in a high curvature region (such as a sharp bend), the dynamic expansion step size decreases to ensure path feasibility.

[0153] The preset step size thresholds include a lower limit and an upper limit for the extended step size, which correspond to the fine exploration step size for narrow curves and the fast exploration step size for wide straight sections, respectively.

[0154] The preset curvature influence factor is used to balance the influence of boundary distance and centerline curvature on the dynamic expansion step size, and to adjust the constraint strength of curvature on the dynamic expansion step size.

[0155] In this step, a dynamically expanding step size is calculated based on the current boundary distance and centerline curvature, realizing a node search strategy based on the actual geometry of the tunnel. Large step sizes are used for rapid traversal and exploration in wide, straight sections, improving path planning efficiency; while in narrow, curved sections, the step size is automatically reduced for refined sampling and searching. This mechanism effectively balances and solves the shortcomings of traditional path planning algorithms in balancing operational efficiency and success rate, ensuring that the overall planning process considers both planning efficiency and path quality.

[0156] In one possible implementation, the preset step size threshold includes a minimum expansion step size and a maximum expansion step size, and a dynamic expansion step size. The calculation formula is as follows:

[0157]

[0158] in, To extend the lower limit of the step size; To extend the upper limit of the step size; For the current node The distance from (referred to as the extension start point in some embodiments) to the nearest tunnel wall (referred to as the boundary line in some embodiments); It is a normalized increasing function, which means that the farther away from the tunnel wall, the larger the available space. The step size can be increased appropriately. For the current node The curvature of the centerline corresponding to the tunnel centerline; It is a normalized decreasing function, which means that the greater the curvature of the centerline, i.e. the sharper the bend, the smaller the step size should be for more precise exploration; A pre-defined curvature influence factor is used to balance the effects of distance and curvature on the step size. By dynamically expanding the step size, the path planning algorithm can quickly explore wide straight sections with larger step sizes, while automatically switching to smaller step sizes for finer searching in narrow curves. More intensive sampling and exploration are performed in curve regions to ensure that feasible paths through narrow curves are found.

[0159] S604. Based on the dynamic expansion step size and the kinematic model of the articulated vehicle, the expansion starting point is expanded along the direction of the direction reference point to generate candidate nodes.

[0160] The articulated vehicle kinematic model is used to ensure that the expansion process satisfies the kinematic constraints of the target articulated vehicle.

[0161] S605. Based on the boundary line and the kinematic model of the articulated vehicle, perform a validity check on the candidate nodes. If the candidate node passes the validity check, proceed to S606; if the candidate node fails the validity check, proceed to S601~S605.

[0162] In this step, the legality verification includes collision detection and articulation angle physical limit check. Specifically, on the one hand, based on the articulated vehicle kinematic model, the complete process of the vehicle moving from the extended starting point to the candidate node is simulated, and the vehicle body envelope pose at different times is calculated. The system checks whether the vehicle body envelope interferes with or collides with the boundary line. If a collision occurs, the candidate node fails the legality verification; if no collision occurs, the candidate node passes the collision detection. On the other hand, based on the articulated vehicle kinematic model, the change in the articulation angle between the front and rear vehicle bodies during the vehicle's movement is calculated. The system checks whether the articulation angle exceeds the vehicle's preset physical limit range. If the articulation angle exceeds the limit, the candidate node fails the legality verification; if the articulation angle does not exceed the limit, the candidate node passes the articulation angle physical limit check. If the candidate node passes both the collision detection and the articulation angle physical limit check, then the candidate node passes the legality verification.

[0163] In one possible implementation, environmental obstacle information on the path from the extended starting point to the candidate node is obtained. When performing legality verification on the candidate node, it is verified whether the vehicle body envelope interferes or collides with the boundary line and environmental obstacles. If a collision occurs, the candidate node is determined to have failed the legality verification.

[0164] S606. The candidate nodes that pass the legality verification are identified as new nodes.

[0165] In this step, the candidate nodes that pass the validity check are associated with their parent-child relationship with the starting point of the expansion, and the new nodes are added to the initial path tree to complete the node addition operation of this iteration expansion.

[0166] This application provides a path planning method for articulated vehicles in tunnels. The method obtains the boundary distance based on the expansion start point and the boundary line, the centerline curvature based on the expansion start point and the tunnel centerline, and the dynamic expansion step size based on the preset step size threshold, the preset curvature influence factor, the boundary distance, and the centerline curvature. The dynamic expansion step size enables adaptive optimization of path search in different environmental regions.

[0167] Subsequently, based on the dynamic expansion step size and the articulated vehicle's kinematic model, the expansion starting point is expanded along the direction of the directional reference point to generate candidate nodes. The candidate nodes are then validated for legitimacy based on the boundary line and the articulated vehicle's kinematic model. The process of obtaining the boundary distance based on the expansion starting point and the boundary line is repeated until a candidate node passes the validity check. The candidate nodes that pass the validity check are then designated as new nodes, thus verifying the feasibility of path planning under physical constraints. Overall, these techniques effectively improve the driving safety of articulated vehicles in narrow tunnel spaces.

[0168] As can be seen from the above embodiments, the path planning method proposed in this application integrates a dual mechanism of tunnel centerline-guided path search and environmentally adaptive dynamic step size expansion. Compared with traditional path planning algorithms, this method can adapt to the restricted scenarios of narrow, winding tunnels with sparse environmental features, and has a higher path planning success rate and environmental robustness. Furthermore, unlike the exhaustive search method that relies on high-precision grid maps, this method adopts a sampling planning approach, coupled with an efficient dynamic expansion step size adaptive adjustment strategy, effectively reducing the overall computational complexity. This design can stably achieve real-time path planning on a vehicle-mounted platform with limited computing power, adapting to the actual driving needs of articulated vehicles in underground tunnels, and possesses outstanding engineering feasibility and practical application value.

[0169] Based on any of the above embodiments, the following, in conjunction with Figure 7 This paper provides a detailed explanation of a path planning method for articulated vehicles in tunnels through specific examples.

[0170] S701, Environment Modeling and Initialization.

[0171] The first step is to construct a point cloud map using laser scanning. A vehicle-mounted LiDAR scanner is used to scan the tunnel environment, acquiring a high-precision set of three-dimensional spatial points to generate a point cloud map of the tunnel environment.

[0172] The second step is to extract tunnel features. A skeleton extraction algorithm is used to extract the continuous centerline and left and right boundary lines of the tunnel from the point cloud map, obtaining the discrete point coordinate sequence of the centerline and the boundary line information.

[0173] The third step is pose initialization. The starting pose (referred to as real-time pose in some embodiments, and specifically the pose of the articulated vehicle when it is at the start of the path) and the ending pose (referred to as target pose in some embodiments) of the articulated vehicle are set to complete the initial parameter configuration for path planning.

[0174] S702, Path Search Loop.

[0175] The path expansion is iteratively performed based on the Improved Rapidly-exploring Random TreeStar (RRT*) algorithm, and the specific process is as follows:

[0176] The first step is random sampling. Based on the RRT* algorithm, direction reference points are randomly sampled in the vehicle's motion state space to provide directional guidance for node expansion.

[0177] The second step is adaptive step size calculation. Based on the direction reference point, the curvature of the expansion starting point and the tunnel centerline, and the boundary distance from the expansion starting point to the tunnel wall (referred to as the boundary line in some embodiments), the dynamic expansion step size for this expansion is obtained, realizing environmentally adaptive node expansion.

[0178] The third step is to expand to new nodes. Based on the expansion starting point and the dynamic expansion step size, candidate nodes are generated.

[0179] Step 4, Feasibility Verification: Perform collision detection (based on the interference judgment between the vehicle envelope and the tunnel boundary line) and hinge angle physical limit check on the candidate nodes. If the verification fails, return to step 1 and iterate again; if the verification passes, proceed to step 5.

[0180] The fifth step is to add the new node to the tree. The new node that passes the verification is added to the current path search tree (referred to as the initial path tree in some embodiments), and the comprehensive cost is calculated using a path planning algorithm to determine the target parent node of the new node, thereby achieving path reconnection optimization.

[0181] Step 6, Target Determination. Check if there is a node in the current path search tree that has entered the neighborhood of the destination pose; if the neighborhood of the destination pose has not been reached, return to step 1 to continue iterating; if the neighborhood of the destination pose has been reached, proceed to step 7.

[0182] Step 7: Extract the path point sequence set. Based on the parent node relationships of each node in the iteratively completed path search tree (referred to as the target path tree in some embodiments), backtrack from the endpoint node entering the endpoint pose neighborhood to the starting node to generate a path point sequence set from the starting pose to the endpoint pose. This path point sequence set is formed by sequentially connecting discrete nodes and is only a feasible global coarse path from the starting point to the endpoint. It cannot directly meet the kinematic requirements for smooth driving of articulated vehicles, so further path smoothing and curvature optimization processing is required.

[0183] S703, Path Smoothing Optimization.

[0184] The first step is piecewise fitting. Fifth-order polynomial splines are used to fit the path point sequence set piecewise.

[0185] The second step is G2 continuity constraint. At the connection points of each curve segment, a forced continuity constraint is applied to ensure the continuity of the position, tangent direction, and curvature of adjacent curves, thus achieving G2 continuity of the curve.

[0186] The third step is optimization. An optimization problem that incorporates the above G2 continuous constraints is constructed and solved, connecting the piecewise smooth curves that satisfy the constraints into a continuous final trajectory.

[0187] S704, Output the final trajectory.

[0188] The final trajectory is output as the driving trajectory of the articulated vehicle adapted to the current tunnel environment, and the process ends. Subsequently, this path will be transmitted to the vehicle's underlying tracking controller to execute control commands such as steering and speed, so as to realize the vehicle's automatic driving.

[0189] It should be noted that, in Figure 7 The processing steps S701-S704 shown in the embodiments do not constitute a specific limitation on a path planning method for articulated vehicles in tunnels. In other embodiments of this application, a path planning method for articulated vehicles in tunnels may include... Figure 7 Embodiments may include more or fewer steps; for example, a path planning method for an articulated vehicle in a tunnel may include... Figure 7 Some steps in the embodiments, or, Figure 7 Some steps in the embodiments can be replaced by steps with the same function, or, Figure 7 Some steps in the embodiments can be broken down into multiple steps, etc.

[0190] Figure 8 A schematic diagram of a path planning device for an articulated vehicle in a tunnel, as provided in this application, is shown below. Figure 8 As shown, the path planning device 80 for an articulated vehicle in a tunnel provided in this embodiment includes:

[0191] The acquisition module 801 is used to acquire the tunnel environment point cloud map of the target tunnel and the real-time pose of the target articulated vehicle.

[0192] The determination module 802 is used to determine the tunnel centerline and boundary line based on the tunnel environment point cloud map.

[0193] The path planning module 803 is used to generate a set of path point sequences based on the real-time pose, boundary line, preset target pose of the target articulated vehicle, and path planning algorithm, using the tunnel centerline as the guiding reference for path planning.

[0194] The generation module 804 is used to generate the driving path of the target articulated vehicle in the target tunnel based on the path point sequence set; the driving path is the path close to the center line of the tunnel in the spatial distribution of the target tunnel.

[0195] In one possible implementation, the path planning module 803 is specifically used for:

[0196] Based on the inherent parameters of the target articulated vehicle, a kinematic model of the articulated vehicle is constructed.

[0197] Based on the tunnel environment point cloud map, real-time pose, and preset target pose, a motion state space is generated.

[0198] Construct an initial path tree using the real-time pose as the root node.

[0199] Using the tunnel centerline as the guiding reference for path planning, the initial path tree is iteratively expanded repeatedly based on the articulated vehicle kinematic model, motion state space, boundary lines, and path planning algorithm until the preset termination conditions are met, thus obtaining the target path tree.

[0200] Based on each node of the target path tree, generate a set of path point sequences.

[0201] In one possible implementation, the path planning module 803 is specifically used for:

[0202] Based on the inherent parameters of the target articulated vehicle, a kinematic model of the articulated vehicle is constructed.

[0203] Based on the tunnel environment point cloud map, real-time pose, and preset target pose, a motion state space is generated.

[0204] Construct an initial path tree using the real-time pose as the root node.

[0205] Randomly sample the motion state space to generate direction reference points.

[0206] Based on the tunnel centerline and boundary line, the comprehensive cost of the path from each existing node in the initial path tree to the direction reference point is calculated using a path planning algorithm, and an extension starting point is selected from each existing node. The comprehensive cost represents the degree to which the path deviates from the tunnel centerline, the degree to which the path is close to the boundary line, and the length of the path.

[0207] Based on the articulated vehicle's kinematic model, direction reference point, tunnel centerline, boundary line, and extension starting point, a new node is generated; the extension starting point is the initial parent node of the new node.

[0208] Based on the tunnel centerline and boundary line, the comprehensive cost of the path from each existing node to the new node is calculated using a path planning algorithm, and the target parent node of the new node is determined from the existing nodes.

[0209] Based on the new node and the target parent node, update the initial path tree; and repeatedly perform the process of randomly sampling the motion state space to generate direction reference points until the preset termination condition is met to obtain the target path tree.

[0210] Based on each node of the target path tree, generate a set of path point sequences.

[0211] In one possible implementation, the path planning module 803 is specifically used for:

[0212] Based on the inherent parameters of the target articulated vehicle, a kinematic model of the articulated vehicle is constructed.

[0213] Based on the tunnel environment point cloud map, real-time pose, and preset target pose, a motion state space is generated.

[0214] Construct an initial path tree using the real-time pose as the root node.

[0215] Randomly sample the motion state space to generate direction reference points.

[0216] Based on the tunnel centerline and boundary line, the comprehensive cost of the path from each existing node in the initial path tree to the direction reference point is calculated using a path planning algorithm, and an extension starting point is selected from each existing node. The comprehensive cost represents the degree to which the path deviates from the tunnel centerline, the degree to which the path is close to the boundary line, and the length of the path. The extension starting point is the initial parent node of the new node.

[0217] Based on the dynamic expansion step size and the articulated vehicle kinematic model, the expansion starting point is expanded along the direction of the direction reference point to generate candidate nodes; the articulated vehicle kinematic model is used to ensure that the expansion process satisfies the kinematic constraints of the target articulated vehicle.

[0218] Based on the boundary line and the kinematic model of the articulated vehicle, the legality of candidate nodes is verified; and the process of dynamically expanding the step size by taking the direction reference point, the tunnel centerline and the boundary line is repeated until the candidate node passes the legality verification.

[0219] Candidate nodes that pass the validity check will be identified as new nodes.

[0220] Based on the tunnel centerline and boundary line, the comprehensive cost of the path from each existing node to the new node is calculated using a path planning algorithm, and the target parent node of the new node is determined from the existing nodes.

[0221] Based on the new node and the target parent node, update the initial path tree; and repeatedly perform the process of randomly sampling the motion state space to generate direction reference points until the preset termination condition is met to obtain the target path tree.

[0222] Based on each node of the target path tree, generate a set of path point sequences.

[0223] In one possible implementation, the path planning module 803 is specifically used for:

[0224] The boundary distance is obtained based on the starting point and boundary line of the extension.

[0225] The curvature of the centerline is obtained based on the starting point of the extension and the centerline of the tunnel.

[0226] The dynamic expansion step size is obtained based on the preset step size threshold, preset curvature influence factor, boundary distance, and centerline curvature. The dynamic expansion step size is positively correlated with the boundary distance and negatively correlated with the centerline curvature.

[0227] In one possible implementation, the generation module 804 is specifically used for:

[0228] The path point sequence set is segmented to obtain multiple path point segment sets and common connection points between each path point segment set.

[0229] Based on the preset path continuity constraints, common connection points, and segmentation sets of each path point, multiple segmented smooth curves are generated.

[0230] The driving trajectory is obtained by connecting the various segmented smooth curves into a continuous whole.

[0231] This embodiment provides a path planning device for articulated vehicles in tunnels, which can execute the method provided in the above-described method embodiments. Its implementation principle and technical effects are similar, and will not be described in detail here.

[0232] Figure 9 A schematic diagram of the structure of the electronic device provided in this application. Figure 9 As shown, the electronic device 90 provided in this embodiment includes at least one processor 901 and a memory 902. Optionally, the device 90 further includes a communication component 903. The processor 901, memory 902, and communication component 903 are connected via a bus 904.

[0233] In a specific implementation, at least one processor 901 executes computer execution instructions stored in memory 902, causing at least one processor 901 to perform the above-described method.

[0234] The specific implementation process of processor 901 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0235] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0236] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0237] 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. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0238] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0239] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described method.

[0240] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0241] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.

[0242] The division of units is merely a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0243] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0244] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0245] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0246] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0247] Finally, it should be noted that other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein, and is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. A path planning method for articulated vehicles in tunnels, characterized in that, include: Acquire the tunnel environment point cloud map of the target tunnel and the real-time pose of the target articulated vehicle; Based on the tunnel environment point cloud map, determine the tunnel centerline and boundary line; Using the tunnel centerline as the guiding reference for path planning, a path point sequence set is generated based on the real-time pose, the boundary line, the preset target pose of the target articulated vehicle, and the path planning algorithm. Based on the path point sequence set, generate the travel path of the target articulated vehicle in the target tunnel; The driving path is a path close to the centerline of the target tunnel in the spatial distribution of the tunnel.

2. The method according to claim 1, characterized in that, The method of using the tunnel centerline as a guiding reference for path planning, and generating a path point sequence set based on the real-time pose, the boundary line, the preset target pose of the target articulated vehicle, and the path planning algorithm, includes: Based on the inherent parameters of the target articulated vehicle, a kinematic model of the articulated vehicle is constructed. Based on the tunnel environment point cloud map, the real-time pose, and the preset target pose, a motion state space is generated; Construct an initial path tree using the real-time pose as the root node; Using the tunnel centerline as the guiding reference for path planning, and based on the articulated vehicle kinematic model, the motion state space, the boundary line, and the path planning algorithm, the initial path tree is iteratively expanded repeatedly until a preset termination condition is met, thereby obtaining the target path tree. Based on each node of the target path tree, the path point sequence set is generated.

3. The method according to claim 2, characterized in that, The method of using the tunnel centerline as the guiding reference for path planning, and based on the articulated vehicle kinematic model, the motion state space, the boundary line, and the path planning algorithm, iteratively expanding the initial path tree until a preset termination condition is met to obtain the target path tree, includes: Randomly sample the motion state space to generate direction reference points; Based on the tunnel centerline and the boundary line, the comprehensive cost of the path from each existing node in the initial path tree to the direction reference point is calculated using the path planning algorithm, and an extension starting point is selected from each of the existing nodes; the comprehensive cost represents the degree to which the path deviates from the tunnel centerline, the degree to which the path is close to the boundary line, and the length of the path; Based on the articulated vehicle kinematic model, the direction reference point, the tunnel centerline, the boundary line, and the extension starting point, a new node is generated; the extension starting point is the initial parent node of the new node. Based on the tunnel centerline and the boundary line, the comprehensive cost of the path from each existing node to the new node is calculated using the path planning algorithm, and the target parent node of the new node is determined from each existing node. The initial path tree is updated based on the new node and the target parent node; and the process of randomly sampling the motion state space and generating direction reference points is repeated until the preset termination condition is met to obtain the target path tree.

4. The method according to claim 3, characterized in that, The process of generating new nodes based on the articulated vehicle kinematic model, the direction reference point, the tunnel centerline, the boundary line, and the extended starting point includes: The dynamic expansion step size is obtained based on the direction reference point, the tunnel centerline, and the boundary line; Based on the dynamic expansion step size and the articulated vehicle kinematic model, the expansion starting point is expanded along the direction of the direction reference point to generate candidate nodes; the articulated vehicle kinematic model is used to ensure that the expansion process satisfies the kinematic constraints of the target articulated vehicle. Based on the boundary line and the articulated vehicle kinematic model, the candidate nodes are validated for legality; and the process of obtaining a dynamic expansion step size based on the direction reference point, the tunnel centerline, and the boundary line is repeated until the candidate nodes pass the legality validation. The candidate nodes that pass the legality verification are determined as the new nodes.

5. The method according to claim 4, characterized in that, The step of obtaining the dynamic expansion step size based on the direction reference point, the tunnel centerline, and the boundary line includes: The boundary distance is obtained based on the extended starting point and the boundary line; The curvature of the centerline is obtained based on the extended starting point and the tunnel centerline; The dynamic expansion step size is obtained based on the preset step size threshold, the preset curvature influence factor, the boundary distance, and the centerline curvature; the dynamic expansion step size is positively correlated with the boundary distance and negatively correlated with the centerline curvature.

6. The method according to claim 1, characterized in that, The step of generating the travel path of the target articulated vehicle in the target tunnel based on the path point sequence set includes: The path point sequence set is segmented to obtain multiple path point segment sets and common connection points between each path point segment set; Based on the preset path continuity constraints, the common connection points, and the segmentation sets of each path point, multiple segmented smooth curves are generated. The driving path is obtained by connecting the segmented smooth curves into a continuous whole.

7. A path planning device for articulated vehicles in tunnels, characterized in that, include: The acquisition module is used to acquire the tunnel environment point cloud map of the target tunnel and the real-time pose of the target articulated vehicle. The determination module is used to determine the tunnel centerline and boundary line based on the tunnel environment point cloud map; The path planning module is used to generate a set of path point sequences based on the real-time pose, the boundary line, the preset target pose of the target articulated vehicle, and the path planning algorithm, using the tunnel centerline as the guiding reference for path planning. The generation module is used to generate the travel path of the target articulated vehicle in the target tunnel based on the path point sequence set; The driving path is a path close to the centerline of the target tunnel in the spatial distribution of the tunnel.

8. An electronic device, characterized in that, include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the processor to perform the method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1 to 6.

10. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method described in any one of claims 1 to 6.

Images