Global and local path layered planning method and device for unstructured roads in mining areas

Abstract

The present invention provides a method and device for hierarchical planning of global and local paths for unstructured roads in mining areas. The method comprises the following steps: obtaining a surveying map of unstructured roads in mining areas to establish a global node-arc model, and planning an initial global path for vehicle travel based on the global node-arc model; splitting the road segments based on the initial nodes, and optimizing and fusing the initial local paths of the entire road segments obtained by splitting to obtain optimized local paths and optimized global paths for the entire road segments; determining the starting and target positions of the vehicle, and determining the optimal global path in the optimized global path, and determining multiple optimal local paths for vehicle tracking based on the optimal global path; and if an obstacle is detected during tracking, re-adjusting the temporary local paths based on preset vehicle kinematics and an artificial potential field method. The method for hierarchical planning of global and local paths for unstructured roads in mining areas provided by the present invention can take into account both the efficiency and effect of path planning and the safety of the planned path.

Description

Technical Field

[0001] The present invention relates to the technical field of path planning under intelligent transportation, and in particular to a method and device for hierarchical planning of global and local paths of unstructured roads in mining areas. Background Art

[0002] In recent years, with the continued development and maturity of autonomous driving applications, path planning for intelligent vehicles has attracted increasing attention. Path planning for autonomous driving is crucial for autonomous vehicle movement and route optimization. Based on actual road map data, intelligent vehicles must find a safe path from their initial location to their target location that minimizes distance or time. This has become a major development trend. Choosing a reasonable, safe, and efficient path is crucial to achieving autonomous driving.

[0003] Most existing global path planning methods perform path search by ellipticalizing the vehicle and then using algorithms such as the A* algorithm, fast search tree algorithm, D* algorithm, and heuristic algorithm. The searched path is then smoothed to arrive at a path suitable for vehicle travel. However, for unstructured roads with special corners such as those in mining areas, on the one hand, each time a global A*-like search algorithm is used to search for a path, the computational complexity is large, the number of search points is large, the search range is wide, and the efficiency is low, and it cannot guarantee the optimal path. On the other hand, the existing real-time local path planning also has algorithmic limitations, and the planned path is not necessarily the optimal path. In addition, the existing path search method will, to a certain extent, cause the redundant space between the vehicle and the walls on both sides to be compressed, resulting in the vehicle's poor ability to avoid risks. Therefore, how to form the safest possible driving path while ensuring the feasibility of the path is an urgent problem that unmanned mining trucks need to solve.

[0004] In summary, the existing path planning technology for unstructured roads has the technical problem of failing to balance planning efficiency, path planning effect, and safety of the planned path. Summary of the Invention

[0005] In view of this, it is necessary to provide a global and local path hierarchical planning method and device for unstructured roads in mining areas to solve the technical problem that the path planning of unstructured roads in the existing technology cannot take into account the planning efficiency, path planning effect and safety of the planned path.

[0006] In order to solve the above technical problems, on the one hand, the present invention provides a global and local path hierarchical planning method for unstructured roads in mining areas, comprising:

[0007] Obtaining a survey map of unstructured roads in the mining area, determining the road boundaries of the unstructured roads, setting initial nodes in key areas to establish a global node-arc model, and planning an initial global path for vehicle traffic based on the global node-arc model;

[0008] Splitting the initial global path into sections according to the initial nodes to obtain initial local paths for the entire section, and optimizing and fusing the initial local paths for the entire section based on vehicle kinematics and redundant space on both sides of the vehicle to obtain optimized local paths and an optimized global path for the entire section;

[0009] Determining a starting position and a target position of the vehicle, determining an optimal global path in the optimized global path according to the starting position and the target position, and determining multiple optimal local paths for the vehicle to follow according to the optimal global path;

[0010] When the vehicle is tracking and driving according to the optimal global path and multiple optimal local paths, and an obstacle appears on the optimal local path, a temporary local path is re-planned based on the preset vehicle kinematics and artificial potential field method, and the vehicle is controlled to track and drive according to the planned temporary local path. After bypassing the obstacle, the vehicle returns to the optimal global path and multiple optimal local paths to continue tracking and driving.

[0011] In some possible implementations, obtaining a survey map of unstructured roads in a mining area, determining the road boundaries of the unstructured roads, setting initial nodes in key areas to establish a global node-arc model, and planning an initial global path for vehicle travel based on the global node-arc model includes:

[0012] By mapping the mining area, a mapping map of the unstructured roads in the mining area is obtained;

[0013] Determine the road boundaries and key areas and curve areas of all roads according to the surveying map, and set initial nodes in the key areas and curve areas;

[0014] All roads included in the surveying and mapping map are used as connecting arcs, the initial node is used as a connecting node, and a node-arc model is constructed;

[0015] An initial global path for vehicle passage is obtained based on the node-arc model.

[0016] In some possible implementations, the splitting of the initial global path into segments according to the initial virtual nodes to obtain initial local paths for the entire segment, and optimizing and fusing the initial local paths for the entire segment based on vehicle kinematics and redundant space on both sides of the vehicle to obtain optimized local paths for the entire segment and an optimized global path include:

[0017] Splitting the entire road segment of the initial global path into straight road segments and curved road segments according to the initial node to obtain the initial local path of the entire road segment;

[0018] Based on the vehicle kinematic model and taking the same redundant space between the vehicle and the walls on both sides as the optimization goal, the straight sections in the initial local path of the entire road section are optimized by ensuring that the vehicle has the same redundant space between the walls on both sides, thereby obtaining an optimized local path for the straight sections;

[0019] determining an optimal state of the vehicle during a turning process based on a relationship between the vehicle's posture and the distance from the inner and outer walls of the road when the vehicle is traveling on the curved road section, taking the optimal state of the vehicle during the turning process as a base condition and the maximum value of the redundant space outside the vehicle as an optimization condition, optimizing the curved section in the initial local path of the entire road section by ensuring that the redundant space inside and outside the vehicle is the same, thereby obtaining an optimized local path for the curved section;

[0020] Based on the difference between the vehicle state and the planned path at the intersection of the trajectories of different sections, the vehicle state in the adjacent sections is used as the basic condition, the vehicle kinematic model is used as the constraint condition, and the planned path difference is zero as the constraint target. The vehicle posture at the intersection of the path trajectories is adjusted, and the intersection of the adjacent sections is optimized to obtain the intersection fusion optimized local path;

[0021] The optimized local path for the entire road section and the optimized global path are obtained according to the optimized local path for the straight road section, the optimized local path for the curved road section and the optimized local path for the intersection and fusion.

[0022] In some possible implementations, determining the optimal state of the vehicle during a turning process based on the relationship between the vehicle's posture and the distance from the inner and outer walls of the road when the vehicle is traveling on the curved road section, taking the optimal state of the vehicle during the turning process as a base condition and the maximum value of the redundant space outside the vehicle as an optimization condition, optimizing the curved section in the initial local path of the entire road section by ensuring the same redundant space inside and outside the vehicle, to obtain an optimized local path for the curved section, includes:

[0023] Acquiring an average curvature of a curved road section and vehicle parameters based on the surveying and mapping map;

[0024] Based on the average curvature of the curved road section, vehicle parameters, and the relationship between the vehicle's posture and the distance from the inner and outer walls of the road when the vehicle is traveling on the curved road section, the distance L from the outer front end of the vehicle to the outer wall is fixed. 外 , forming different vehicle trajectories and solving different L 外 The shortest distance L between the inner side of the lower vehicle and the inner wall 内 , when the L 外 With L 内If they are equal, the current vehicle state is determined to be the optimal state of the vehicle during the turning process, and the optimal state of the vehicle is used as the middle position of the optimal path during the turning process;

[0025] Using the vehicle kinematic model as a constraint, the optimal state of the vehicle as a basic condition, the increment of the front wheel steering angle as a dynamic constraint, and the maximum redundant space outside the vehicle as the optimization condition, the optimal solution is when the redundant space inside and outside the vehicle is consistent. The optimal path for the first half and the optimal path for the second half of the middle position of the curved road section are determined to obtain the optimized local path of the curved road section;

[0026] The initial node of each curved road section is re-determined on the optimized vehicle trajectory to obtain the starting node and ending node of each curved road section.

[0027] In some possible implementations, based on the difference between the vehicle state and the planned path at the intersection of trajectories of different road sections, the vehicle state in adjacent road sections is used as a basic condition, the vehicle kinematic model is used as a constraint condition, and the planned path difference is zero as a constraint target. The vehicle posture and the planned trajectory at the intersection of the path trajectories are adjusted, and the intersection of the adjacent road sections is optimized to obtain the intersection fusion optimized local path, including:

[0028] Obtain different vehicle states and planned trajectories of adjacent straight road segments and curved road segments at the starting and ending nodes of the curved road segments;

[0029] Based on the different vehicle states as the basic conditions, the vehicle kinematic model as the constraint condition, and the zero difference between the planned paths at the starting node and the ending node as the constraint target, the vehicle posture and planned trajectory in the straight section are adjusted so that the trajectories at the starting node and the ending node of the curved road section form a smooth transition curve with the planned trajectory of the straight section, thereby obtaining an intersection and fusion optimized local path.

[0030] In some possible implementations, determining a starting position and a target position of the vehicle, determining an optimal global path in the optimized global path based on the starting position and the target position, and determining multiple optimal local paths for the vehicle to follow based on the optimal global path include:

[0031] Get the starting position and target position of the vehicle;

[0032] Determining an optimal global path of a node-arc model from a starting position to a target position in the node-arc model according to a preset global path search algorithm;

[0033] Determine an optimal global path based on the optimal global path of the node-arc model, and determine multiple local optimal paths based on the optimal global path;

[0034] The vehicle tracks and drives according to the multiple local optimal paths and the optimal global path.

[0035] In some possible implementations, the vehicle tracks and drives along the optimal global path and the multiple optimal local paths. If an obstacle is detected on the optimal local path, a temporary local path is replanned based on preset vehicle kinematics and an artificial potential field method. The vehicle tracks and drives along the planned temporary local path until the obstacle is bypassed and the vehicle returns to the optimal global path and the multiple optimal local paths to continue tracking, including:

[0036] The vehicle tracks and drives from the starting position to the end position according to the preset optimal global path and multiple optimal local paths;

[0037] When the vehicle sensor senses an obstacle on the optimal local path, determining whether the vehicle can safely pass the obstacle based on road conditions;

[0038] When the road conditions allow the vehicle to safely pass the obstacle, a temporary local path is replanned based on the preset vehicle kinematics and artificial potential field method. The vehicle tracks and drives along the planned temporary local path, bypasses the obstacle, and returns to the optimal global path and multiple optimal local paths to continue tracking.

[0039] When the road conditions do not allow the vehicle to safely pass the obstacle, the vehicle returns to the previous node of the current path, takes the previous node as the starting point, and replans the path using the Dijkstra algorithm. The vehicle continues to track according to the obtained temporary optimal global path and temporary multi-segment optimal local paths.

[0040] In some possible implementations, the re-performing of temporary local path planning based on preset vehicle kinematics and artificial potential field method includes:

[0041] In the process of temporary local path planning using the artificial potential field method, when the vehicle passes through the obstacle, there is an attraction force of the next node on the vehicle, a repulsion force of the obstacle on the vehicle, and a repulsion force from the boundary of the vehicle. When the combined force of the attraction, repulsion, and boundary repulsion is zero, the vehicle is trapped in a local minimum point and cannot move.

[0042] Based on the artificial potential field method, an escape force is added to the vehicle to make the vehicle move out of the local minimum point. The direction of the escape force is perpendicular to the attraction of the next node and points to the inside of the road.

[0043] The escape force gradually increases from zero until the vehicle is driven out of the minimum point, at which point the escape force disappears and the vehicle continues to travel according to the current potential field.

[0044] In some possible implementations, determining the vehicle boundary repulsion includes:

[0045] By establishing a road boundary repulsive potential field, the vehicle is prevented from colliding with the lane boundary during the obstacle avoidance process. The specific formula for establishing the lane boundary repulsive potential field is:

[0046]

[0047] Among them, F rep.bo is the road boundary repulsion, η rep.bo is the proportional gain coefficient, e is a natural constant, s is the shortest distance from the vehicle's center of mass to the global optimal path and the local optimal path, and D is the current lane width when the vehicle is running.

[0048] On the other hand, the present invention also provides a device for hierarchical planning of global and local paths for unstructured roads in mining areas, comprising:

[0049] a map model building module, configured to obtain a survey map of unstructured roads in a mining area, determine the road boundaries of the unstructured roads, set initial nodes in key areas to establish a global node-arc model, and plan an initial global path for vehicle traffic based on the global node-arc model;

[0050] a path optimization fusion module, configured to split the initial global path into sections according to the initial nodes to obtain initial local paths for the entire section, and optimize and fuse the initial local paths for the entire section based on vehicle kinematics and redundant space on both sides of the vehicle to obtain optimized local paths for the entire section and an optimized global path;

[0051] A path planning and tracking module is used to determine the starting position and target position of the vehicle, determine the optimal global path in the optimized global path based on the starting and target positions, and determine multiple optimal local paths for the vehicle to track based on the optimal global path;

[0052] The temporary planning and tracking module is used to, when the vehicle is tracking and driving according to the optimal global path and the multiple optimal local paths, and an obstacle appears on the optimal local path, re-plan the temporary local path based on the preset vehicle kinematics and artificial potential field method, control the vehicle to track and drive according to the planned temporary local path, and return to the optimal global path and the multiple optimal local paths to continue tracking and driving after bypassing the obstacle.

[0053] The beneficial effects of adopting the above embodiment are as follows: the global and local path hierarchical planning method for unstructured roads in mining areas provided by the present invention, on the one hand, establishes a global path by establishing a node-arc model, and determines the initial global path by adopting a redundant space optimization method for the path between nodes. By establishing the node-arc model, the calculation amount is smaller, the efficiency is higher, and the path planning effect is better when performing path search planning; on the other hand, the node-arc model is used to combine the vehicle kinematic model during the path optimization process, so that the vehicle path requires fewer nodes to be smoothed, the vehicle driving stability is enhanced, and the safety of the planned path is increased; on the other hand, by setting the redundant space distance between the vehicle and the walls on both sides as the optimization condition, the maximum safe distance can be maintained between the vehicle and the walls on both sides, further ensuring the safety of vehicle driving; finally, by setting the improved artificial potential field method, the vehicle has the ability to plan a local obstacle avoidance path, and the Dijkstra algorithm is used to re-plan the path, which improves the vehicle's emergency avoidance capability, while increasing traffic safety, it also ensures the efficiency of the re-planned path to reach the target location. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0055] Figure 1 A schematic flow chart of an embodiment of the method for hierarchical planning of global and local paths for unstructured roads in mining areas provided by the present invention;

[0056] Figure 2 The present invention provides Figure 1 A flow chart of an embodiment of step S101;

[0057] Figure 3 The present invention provides Figure 1 A flow chart of an embodiment of step S102;

[0058] Figure 4 A schematic diagram of an embodiment of the curve path optimization provided by the present invention;

[0059] Figure 5 A schematic diagram of an embodiment of path fusion provided by the present invention;

[0060] Figure 6 A schematic diagram of an embodiment of a mine road planning route provided by the present invention;

[0061] Figure 7 The present invention provides Figure 1 A flow chart of an embodiment of step S103;

[0062] Figure 8 The present invention provides Figure 1 A flow chart of an embodiment of step S104;

[0063] Figure 9 A schematic diagram of an embodiment of obstacle avoidance path planning provided by the present invention;

[0064] Figure 10 This is a structural diagram of an embodiment of a device for hierarchical planning of global and local paths for unstructured roads in mining areas provided by the present invention. DETAILED DESCRIPTION

[0065] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts shall fall within the scope of protection of the present invention.

[0066] It should be understood that the schematic drawings are not drawn to scale. The flowcharts used in the present invention illustrate operations implemented according to some embodiments of the present invention. It should be understood that the operations in the flowcharts may be implemented out of sequence, and steps that do not have a logical contextual relationship may be reversed or performed simultaneously. In addition, those skilled in the art, guided by the present disclosure, may add one or more additional operations to the flowcharts or remove one or more operations from the flowcharts.

[0067] Some of the blocks shown in the accompanying drawings are functional entities that do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different networks and / or processor systems and / or microcontroller systems.

[0068] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present invention. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute a separate or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0069] Before describing the specific embodiments, the terms used in the present invention are explained:

[0070] Artificial Potential Field Method: This is a classic path planning algorithm that treats the target and obstacles as objects that exert gravitational and repulsive forces on the robot, respectively. The robot moves along the combined force of gravitational and repulsive forces.

[0071] A* Algorithm: The A* (A-Star) algorithm is a highly effective direct search method for finding the shortest path in static road networks. It is also an effective algorithm for solving many search problems. The closer the distance estimate in the algorithm is to the actual value, the faster the search will be.

[0072] D* Algorithm: D* is a dynamic A* algorithm, developed from the static A* algorithm. It is a heuristic path search algorithm suitable for scenarios with unknown or dynamically changing surrounding environments. Unlike A, which searches from the starting point to the target point, D* searches from the target point back to the starting point, using reverse propagation.

[0073] Dijkstra's algorithm: Proposed by Dutch computer scientist Dijkstra in 1959, it's also known as the Dijkstra algorithm. It's a shortest path algorithm from one vertex to all other vertices, solving the shortest path problem in weighted graphs. The key feature of the Dijkstra algorithm is that it starts from a starting point and uses a greedy algorithm to traverse the nearest unvisited neighboring nodes until it reaches the end point.

[0074] Based on the above-mentioned term description, the existing technology uses search algorithms to plan paths for unstructured roads, but the path search often involves large amounts of computation, a large number of search points, a large search range, and low efficiency, resulting in the inability to guarantee the efficiency and optimal path of path planning. In addition, traditional path planning methods for unstructured roads do not consider the impact of road structure on driving safety when planning paths, resulting in the planned path being unable to guarantee vehicle driving safety. Therefore, this invention aims to propose a global and local path hierarchical planning method for unstructured roads in mining areas that can take into account planning efficiency, path planning effect, and planned path safety.

[0075] The embodiments of the present invention provide a method and device for hierarchical planning of global and local paths for unstructured roads in a mining area, which are described below respectively.

[0076] like Figure 1 As shown, Figure 1 This is a flow chart of an embodiment of a method for hierarchical planning of global and local paths for unstructured roads in mining areas provided by the present invention. The method for hierarchical planning of global and local paths for unstructured roads in mining areas includes:

[0077] S101, obtaining a surveying map of unstructured roads in a mining area, determining the road boundaries of the unstructured roads, setting initial nodes in key areas to establish a global node-arc model, and planning an initial global path for vehicle traffic based on the global node-arc model;

[0078] S102, splitting the initial global path into sections according to the initial nodes to obtain initial local paths for the entire section, and optimizing and fusing the initial local paths for the entire section based on vehicle kinematics and redundant space on both sides of the vehicle to obtain optimized local paths and an optimized global path for the entire section;

[0079] S103, determining a starting position and a target position of the vehicle, determining an optimal global path in the optimized global path based on the starting position and the target position, and determining multiple optimal local paths for the vehicle to follow based on the optimal global path;

[0080] S104. When the vehicle is tracking and driving according to the optimal global path and the multiple optimal local paths, and an obstacle appears on the optimal local path, a temporary local path is re-planned based on the preset vehicle kinematics and artificial potential field method, and the vehicle is controlled to track and drive according to the planned temporary local path. After bypassing the obstacle, the vehicle returns to the optimal global path and the multiple optimal local paths to continue tracking and driving.

[0081] Compared with the prior art, the global and local path hierarchical planning method for unstructured roads in mining areas provided by an embodiment of the present invention, on the one hand, establishes a global path by establishing a node-arc model, and determines the optimized global path by using a redundant space optimization method for the path between nodes. By establishing a node-arc model, the calculation amount during path search planning is reduced, the efficiency is higher, and the path planning effect is better; on the other hand, the node-arc model is used to combine the vehicle kinematic model during path optimization, so that the vehicle path has fewer nodes that need to be smoothed, the vehicle driving stability is enhanced, and the safety of the planned path is increased; on the other hand, by setting the distance of the redundant space between the vehicle and the walls on both sides as the optimization condition, the vehicle can maintain a maximum safe distance from the walls on both sides, further ensuring the safety of vehicle driving; finally, by setting an improved artificial potential field method, the vehicle has the ability to plan a local obstacle avoidance path, thereby improving the vehicle's emergency risk avoidance capability.

[0082] In some embodiments of the present invention, Figure 2 As shown, Figure 2 The present invention provides Figure 1 The flowchart of step S101 in an embodiment of the present invention is as follows: step S101 includes:

[0083] S201, obtaining a survey map of unstructured roads in the mining area by surveying and mapping the mining area;

[0084] S202: determining the road boundaries, key areas, and curve areas of all roads according to the surveying map, and setting initial nodes in the key areas and curve areas;

[0085] S203, constructing a node-arc model using all roads included in the surveying and mapping map as connecting arcs and the initial node as a connecting node;

[0086] S204: Acquire an initial global path available for vehicle passage based on the node-arc model.

[0087] In a specific embodiment of the present invention, a surveying map of the mining area boundary is obtained by surveying the mining area, and virtual nodes are set in key areas of the mining area roads (mainly including bends and other non-road areas). Based on these virtual nodes, the roads are used as connecting arcs, and the virtual nodes are used as connecting nodes to establish the initial global path of the node-arc model. All virtual nodes are initial nodes, and the initial global path of the surveying map is established through the path nodes.

[0088] The embodiment of the present invention establishes a global path by establishing a node-arc model, simplifying the key areas and roads in the unstructured roads in the mining area into virtual nodes and connecting arcs, so that the calculation amount during path search planning is smaller, the efficiency is higher, and the path planning effect is better.

[0089] Furthermore, in some embodiments of the present invention, Figure 3 As shown, Figure 3 The present invention provides Figure 1 The flowchart of step S102 of an embodiment is shown in FIG. 1 , wherein step S102 includes:

[0090] S301, splitting the entire road segment of the initial global path into straight road segments and curved road segments according to the initial node, to obtain the initial local path of the entire road segment;

[0091] S302: Based on the vehicle kinematic model and taking the redundant space between the vehicle and the walls on both sides as the optimization goal, the straight sections in the initial partial path of the entire road section are optimized by ensuring that the redundant space between the vehicle and the walls on both sides is the same, thereby obtaining an optimized partial path for the straight sections;

[0092] S303: Determine an optimal state of the vehicle during a turning process based on a relationship between the vehicle's posture and the distance from the inner and outer walls of the road when the vehicle is traveling on the curved road section; using the optimal state of the vehicle during the turning process as a base condition and the maximum value of the redundant space outside the vehicle as an optimization condition, optimize the curved section in the initial partial path of the entire road section by ensuring that the redundant space inside and outside the vehicle is the same, thereby obtaining an optimized partial path for the curved section;

[0093] S304: Based on the difference between the vehicle state and the planned path at the intersection of the trajectories of different road sections, the vehicle state in the adjacent road sections is used as a basic condition, the vehicle kinematic model is used as a constraint condition, and the planned path difference is zero as a constraint target. The intersection of the adjacent road sections is optimized to obtain an intersection fusion optimized local path.

[0094] S305 , obtaining the optimized local path for the entire road section and the optimized global path according to the optimized local path for the straight road section, the optimized local path for the curved road section, and the optimized local path for the intersection and fusion.

[0095] In a specific embodiment of the present invention, in step S301, since the posture of the vehicle is different in straight sections and curved sections, the distances between different parts of the vehicle body and the walls on both sides of the road in the curved sections are not the same. Conventional path planning can easily cause the vehicle to collide with both sides of the road. Therefore, it is necessary to split and optimize the planned initial path. The entire section of the initial global path is split into straight sections and curved sections according to the location of the initial nodes to obtain the initial local path of the entire section.

[0096] Furthermore, in step S302, the path of the straight road section is optimized. Based on the vehicle kinematic model, the optimization goal is to ensure that the redundant space between the vehicle and the walls on both sides of the road is the same. By ensuring that the vehicle has the same redundant space between the walls on both sides, the straight road section in the initial local path of the entire road section is optimized to obtain the optimized local path of the straight road section. The vehicle kinematic model is as follows:

[0097]

[0098]

[0099]

[0100]

[0101]

[0102] in, is the x-direction position of the vehicle's center of mass at the next moment; is the y-direction position of the vehicle's center of mass at the next moment; is the vehicle speed at the next moment; is the yaw angle of the vehicle in the next moment; β is the sideslip angle of the center of mass; the speed in the mining area is slow, so the β value is small and can be ignored; δ f is the front wheel turning angle; δ r is the rear wheel turning angle; l f is the distance between the front axle and the center of mass; l r is the distance between the rear axle and the center of mass.

[0103] Specifically, based on the global path of the node-arc model, the path optimization of the straight road segment is performed. The vehicle kinematic model is used as a constraint, and the redundant space between the vehicle and the two side walls is used as the optimization target. By ensuring the same distance between the vehicle and the two side walls, that is, ensuring the same redundant space as a constraint, the optimal local path between each node is formed. In this embodiment, the positional relationship between the vehicle and the two side walls in the trajectory optimization of the inter-node segment satisfies:

[0104] L 左 =L 右

[0105] Among them L 左 Indicates the shortest distance between the left side of the vehicle and the wall; L 右 The distance between the right side of the vehicle and the wall is the shortest distance. When the above constraints are met at any moment while the vehicle is traveling on a straight road segment, the vehicle can always maintain sufficient redundant space on both sides for emergency maneuvers when traveling on the local road segment, thereby forming an optimized local path for the vehicle on the straight road segment.

[0106] Furthermore, in step S303, the path of the curved road section is optimized, including:

[0107] Acquiring an average curvature of a curved road section and vehicle parameters based on the surveying and mapping map;

[0108] Based on the average curvature of the curved road section, vehicle parameters, and the relationship between the vehicle's posture and the distance from the inner and outer walls of the road when the vehicle is traveling on the curved road section, the distance L from the outer front end of the vehicle to the outer wall is fixed. 外 , forming different vehicle trajectories and solving different L 外 The shortest distance L between the inner side of the lower vehicle and the inner wall 内 , when the L 外 With L 内 If they are equal, the current vehicle state is determined to be the optimal state of the vehicle during the turning process, and the optimal state of the vehicle is used as the middle position of the optimal path during the turning process;

[0109] Using the vehicle kinematic model as a constraint, the optimal state of the vehicle as a basic condition, the increment of the front wheel steering angle as a dynamic constraint, and the maximum redundant space outside the vehicle as the optimization condition, the optimal solution is when the redundant space inside and outside the vehicle is consistent. The optimal path for the first half and the optimal path for the second half of the middle position of the curved road section are determined to obtain the optimized local path of the curved road section;

[0110] The initial node of each curved road section is re-determined on the optimized vehicle trajectory to obtain the starting node and ending node of each curved road section.

[0111] Specifically, for a road section with a large curvature such as a curve, the optimization method for the local section is as follows: Figure 4 As shown, Figure 4 Schematic diagram of an embodiment of the curve path optimization provided by the present invention, L 内 Indicates the shortest distance between the vehicle and the inner wall; L 外 Indicates the shortest distance between the vehicle and the outer wall. Due to the difference between the inner and outer wheels when the vehicle turns, compared with L 内 The value of L 外 The value is easier to measure by fixing L 外 The vehicle kinematics model is used to determine a complete vehicle turning trajectory and calculate the L 外 Based on the front wheel angular range as a kinematic constraint, we can get any possible position of the vehicle at the next moment. Based on the next vehicle position, we can get the distance L between the outside of the vehicle and the wall. 外 , the shortest distance L between the inner side of the vehicle and the wall 内 , the evaluation index of the optimal solution is: θ

[0112]

[0113] The position where the θ value is closest to 0.5 is the optimal solution position for the next moment. Figure 4 The black solid rectangle represents the optimal vehicle state.

[0114] By adjusting L 外 The value of L can form multiple vehicle turning trajectories. 外 The increase of value, minL 内 Gradually decreases when the following conditions are met:

[0115] max(L 外 )=min(L 内 )

[0116] At this time, the vehicle's internal and external redundant spaces are both in their largest overall state, and the vehicle state at this time is taken as the optimal vehicle state during the turning process.

[0117] Based on the maximum continuous increment of the front wheel angle, with the optimal state of the vehicle as the most basic condition, considering vehicle posture factors including speed, wheel angle, vehicle yaw angle, etc., all possible states of the vehicle at the previous moment or the next moment are calculated, and the redundant space on the inside and outside of the optimal vehicle under each vehicle state is calculated. The optimal vehicle state is selected as the basic condition for calculating the vehicle state at the next moment, and finally the optimal driving trajectory of the vehicle in the entire local path of the curve is obtained, and the starting node and end node of each curve section are obtained to replace the original initial node.

[0118] Furthermore, in step S304, after completing the path optimization for the straight and curved sections, the overall planned path obtained is not continuous, and the vehicle states are also different. Therefore, it is necessary to fuse the partial paths at the junctions of the straight and curved sections to make the overall planned path continuous. By obtaining the different vehicle states and planned trajectories at the start and end nodes of the adjacent straight and curved sections; using the different vehicle states as the basis, the vehicle kinematic model as the constraint condition, and the zero difference in the planned paths at the start and end nodes as the constraint target, the vehicle posture and planned trajectory in the straight section are adjusted so that the trajectories at the start and end nodes of the curved section form a smooth transition curve with the planned trajectory of the straight section, thereby obtaining a connecting and fusion optimized local path.

[0119] Specifically, after completing the local path optimization between all nodes, there will be differences in the vehicle states at the intersection of the local path trajectories. Therefore, it is necessary to smooth the optimal paths in adjacent sections to form a relatively smooth transition path, such as Figure 5 As shown in the figure, Figure 5 This is a schematic diagram of an embodiment of the path fusion provided by the present invention, wherein the dotted box represents the fused vehicle posture adjustment, and the fusion points are all new nodes after the curve optimization path, so that the optimal paths between road sections can be fused to form a fused path.

[0120] Furthermore, in step S305, after the path fusion is completed, the optimized local path of the entire road section and the optimized global path are obtained based on the optimized local path of the straight road section, the optimized local path of the curved road section, and the optimized local path of the intersection fusion.

[0121] Specifically, after completing the smoothing of the local optimal paths between all nodes, the mine area unmanned truck path planning based on the surveying map generates the offline mine area unmanned vehicle full road section optimized local path and optimized global path. Figure 6 , Figure 6 This is a schematic diagram of an embodiment of a mine road planning path provided by the present invention. The vehicle travels from the starting position of the loading area to the target position of the unloading area. The starting position and target position of the mine car are marked. The black dots on the road section are all road nodes. The black thin solid line paths between the nodes are all local optimal paths. The local paths between all the nodes finally determined are processed offline; the Dijkstra algorithm is used to search for the global optimal path from the starting position to the target position in the global node-arc model. The actual path length between adjacent nodes is used as the cost estimate. The total cost is minimized (the shortest vehicle travel distance) is optimized to find the global optimal path in the global path. It is represented by the black thick solid line path in the figure. The preset local optimal path on the path is called, and the vehicle travels to the target position according to the global optimal path + the preset local optimal path.

[0122] The embodiment of the present invention uses a node-arc model in conjunction with a vehicle kinematic model during path optimization. This reduces the number of nodes in the vehicle path that require smoothing, enhances vehicle driving stability, and increases the safety of the planned path. By setting the redundant space distance between the vehicle and the walls on both sides as an optimization condition, the vehicle can maintain a maximum safe distance from the walls on both sides, further ensuring vehicle driving safety.

[0123] Furthermore, in some embodiments of the present invention, Figure 7 As shown, Figure 7 The present invention provides Figure 1 The flowchart of step S103 in the embodiment of FIG. 1 is shown, and step S103 includes:

[0124] S701, obtaining the starting position and target position of the vehicle;

[0125] S702: determining an optimal global path of the node-arc model from a starting position to a target position in the node-arc model according to a preset global path search algorithm;

[0126] S703, determining an optimal global path according to the optimal global path of the node-arc model, and determining multiple local optimal paths according to the optimal global path;

[0127] S704: The vehicle tracks and drives according to the multiple local optimal paths and the optimal global path.

[0128] In a specific embodiment of the present invention, the optimized local and global paths for the entire road segment, obtained through path optimization, are stored offline and used on transport vehicles. This allows for application beyond existing work vehicles, simplifying path updates after switching to a different vehicle, and providing strong adaptability. By obtaining the vehicle's starting and destination locations, a global path search algorithm is employed to retrieve the optimal global path from the offline stored node-arc model. Based on this optimal global path, the local optimal path for vehicle tracking is obtained.

[0129] Furthermore, in some embodiments of the present invention, Figure 8 As shown, Figure 8 The present invention provides Figure 1 Schematic diagram of a flow chart of an embodiment of step S104, step S104 includes:

[0130] S801: The vehicle tracks and drives from the starting position to the end position according to the preset optimal global path and multiple optimal local paths;

[0131] S802: When a vehicle sensor senses an obstacle on the optimal local path, determining whether the vehicle can safely pass through the obstacle based on road conditions;

[0132] S803. When the road conditions allow the vehicle to safely pass the obstacle, a temporary local path is replanned based on the preset vehicle kinematics and artificial potential field method. The vehicle follows the planned temporary local path, bypasses the obstacle, and returns to the optimal global path and the multiple optimal local paths to continue following the path.

[0133] S804. When the road conditions do not allow the vehicle to safely pass the obstacle, the vehicle returns to the previous node of the current path, takes the previous node as the starting point, and replans the path using the Dijkstra algorithm. The vehicle continues to track and drive based on the obtained temporary optimal global path and temporary multi-segment optimal local paths.

[0134] In a specific embodiment of the present invention, when a vehicle is operating, it tracks and drives from a starting position to a terminal position according to a preset optimal global path and multiple optimal local paths. During the driving process, the vehicle's on-board radar and other sensors perceive the surrounding environment in real time. When it is detected that the road environment in the current path is different from the preset local path environment or an obstacle appears, it is determined whether the vehicle can safely pass the obstacle based on the road conditions. When the road conditions allow the vehicle to safely pass the obstacle, a temporary local path planning (temporary local path planning uses a potential field method and also follows the constraints of vehicle kinematics and turning radius) is used to perform local obstacle avoidance, such as Figure 9 As shown, Figure 9 This is a schematic diagram of an embodiment of the obstacle avoidance path planning provided by the present invention. After obstacle avoidance is completed, the vehicle returns to the preset optimal global path and multiple optimal local paths and continues to move forward. When the road conditions do not allow the vehicle to safely pass the obstacle, the vehicle returns to the previous node of the current path, takes the previous node as the starting point, and re-plans the path using the Dijkstra algorithm. The vehicle continues to track and drive based on the obtained temporary optimal global path and temporary multiple optimal local paths.

[0135] Specifically, an improved artificial potential field method is used for local obstacle avoidance. The target point is the next node of the road section. In view of the fact that the mine car may collide with the lane boundary during the obstacle avoidance process, a road boundary repulsive potential field is established to prevent the vehicle from colliding with the lane boundary during the obstacle avoidance process. The specific formula for establishing the lane boundary repulsive potential field is:

[0136]

[0137] Among them, F rep.bo is the road boundary repulsion, η rep.bois the proportional gain coefficient, e is a natural constant, s is the shortest distance from the vehicle's center of mass to the global optimal path and the local optimal path, and D is the current lane width during vehicle operation. Combining this with the repulsive and attractive forces of the traditional artificial potential field method creates a new combined force that guides the mining truck in temporary local path planning.

[0138] In the process of local obstacle avoidance, there may be a situation where the attraction of the next node to the vehicle, the repulsion of the obstacle to the vehicle, and the resultant force of the vehicle boundary force are zero. When the resultant force is zero, the vehicle falls into the local minimum point and cannot move. At this time, an external force needs to be added to intervene, which is the escape force F. esc , forcing the vehicle to drive out of the local minimum point. The escape force direction is perpendicular to the attraction of the next node and points to the inner side of the road. The escape force is set to gradually increase from zero to an appropriate value and stop until the vehicle is driven out of the minimum point. The escape force disappears and the vehicle continues to drive according to the current potential field.

[0139] In order to enable the vehicle to return to the optimal global path and multiple optimal local paths as quickly as possible after avoiding obstacles, the gravitational potential field of the preset local optimal path is added. When the vehicle avoids the obstacle, the gravitational potential field of the preset local optimal path takes effect, allowing the vehicle to quickly return to the preset local optimal path. When the vehicle returns to the preset local optimal path, the temporary local path planning ends and the vehicle continues to track and drive according to the optimal global path and multiple optimal local paths.

[0140] The embodiment of the present invention sets an improved artificial potential field method to enable the vehicle to have the ability to plan a local obstacle avoidance path, and adopts the Dijkstra algorithm to re-plan the path, thereby improving the vehicle's emergency avoidance capability. While increasing traffic safety, it also ensures the efficiency of the re-planned path to reach the target location.

[0141] In order to better implement the global and local path hierarchical planning method for unstructured roads in mining areas in the embodiment of the present invention, based on the global and local path hierarchical planning method for unstructured roads in mining areas, the embodiment of the present invention also provides a global and local path hierarchical planning device for unstructured roads in mining areas, such as Figure 10 As shown, the global and local path hierarchical planning device 1000 for unstructured roads in a mining area includes:

[0142] A map model building module 1001 is configured to obtain a survey map of unstructured roads in a mining area, determine the road boundaries of the unstructured roads, set initial nodes in key areas, establish a global node-arc model, and plan an initial global path for vehicle traffic based on the global node-arc model.

[0143] A path optimization fusion module 1002 is configured to split the initial global path into sections according to the initial nodes to obtain initial local paths for the entire section, and optimize and fuse the initial local paths for the entire section based on vehicle kinematics and redundant space on both sides of the vehicle to obtain optimized local paths for the entire section and an optimized global path;

[0144] The path planning and tracking module 1003 is used to determine the starting position and the target position of the vehicle, determine the optimal global path in the optimized global path based on the starting and target positions, and determine multiple optimal local paths for the vehicle to track based on the optimal global path;

[0145] The temporary planning and tracking module 1004 is used to, when the vehicle is tracking and driving according to the optimal global path and the multiple optimal local paths, and an obstacle appears on the optimal local path, re-plan the temporary local path based on the preset vehicle kinematics and artificial potential field method, control the vehicle to track and drive according to the planned temporary local path, and return to the optimal global path and the multiple optimal local paths to continue tracking and driving after bypassing the obstacle.

[0146] The global and local path hierarchical planning device 1000 for unstructured roads in mining areas provided in the above embodiment can implement the technical solution described in the above embodiment of the global and local path hierarchical planning method for unstructured roads in mining areas. The specific implementation principles of the above modules or units can be found in the corresponding contents in the above embodiment of the global and local path hierarchical planning method for unstructured roads in mining areas, and will not be repeated here.

[0147] Those skilled in the art will appreciate that all or part of the process steps of the above-described embodiments can be implemented by instructing related hardware (such as a processor, a controller, etc.) through a computer program, and the computer program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a magnetic disk, an optical disk, a read-only memory, or a random access memory.

[0148] The above is a detailed introduction to the global and local path hierarchical planning method and device for unstructured roads in mining areas provided by the present invention. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for technical personnel in this field, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A global and local path hierarchical planning method for unstructured roads in mining areas, characterized in that: include: Obtaining a survey map of unstructured roads in the mining area, determining the road boundaries of the unstructured roads, setting initial nodes in key areas to establish a global node-arc model, and planning an initial global path for vehicle traffic based on the global node-arc model; Splitting the initial global path into sections according to the initial nodes to obtain initial local paths for the entire section, and optimizing and fusing the initial local paths for the entire section based on vehicle kinematics and redundant space on both sides of the vehicle to obtain optimized local paths and an optimized global path for the entire section; Determining a starting position and a target position of the vehicle, determining an optimal global path in the optimized global path according to the starting position and the target position, and determining multiple optimal local paths for the vehicle to follow according to the optimal global path; When the vehicle is tracking and driving according to the optimal global path and the multiple optimal local paths, and an obstacle appears on the optimal local path, a temporary local path is re-planned based on the preset vehicle kinematics and artificial potential field method, and the vehicle is controlled to track and drive according to the planned temporary local path. After bypassing the obstacle, the vehicle returns to the optimal global path and the multiple optimal local paths to continue tracking and driving; The step of splitting the initial global path into sections according to the initial nodes to obtain initial local paths for the entire section, and optimizing and fusing the initial local paths for the entire section based on vehicle kinematics and redundant space on both sides of the vehicle to obtain optimized local paths and an optimized global path for the entire section includes: Splitting the entire road segment of the initial global path into straight road segments and curved road segments according to the initial node to obtain the initial local path of the entire road segment; Based on the vehicle kinematic model and taking the same redundant space between the vehicle and the walls on both sides as the optimization goal, the straight sections in the initial local path of the entire road section are optimized by ensuring that the vehicle has the same redundant space between the walls on both sides, thereby obtaining an optimized local path for the straight sections; determining an optimal state of the vehicle during a turning process based on a relationship between the vehicle's posture and the distance from the inner and outer walls of the road when the vehicle is traveling on the curved road section, taking the optimal state of the vehicle during the turning process as a base condition and the maximum value of the redundant space outside the vehicle as an optimization condition, optimizing the curved section in the initial local path of the entire road section by ensuring that the redundant space inside and outside the vehicle is the same, thereby obtaining an optimized local path for the curved section; Based on the difference between the vehicle state and the planned path at the intersection of the trajectories of different sections, the vehicle state in the adjacent sections is used as the basic condition, the vehicle kinematic model is used as the constraint condition, and the planned path difference is zero as the constraint target. The vehicle posture at the intersection of the path trajectories is adjusted, and the intersection of the adjacent sections is optimized to obtain the intersection fusion optimized local path; The optimized local path for the entire road section and the optimized global path are obtained according to the optimized local path for the straight road section, the optimized local path for the curved road section and the optimized local path for the intersection and fusion.

2. The method for global and local path hierarchical planning of unstructured roads in mining areas according to claim 1 is characterized in that: The method includes obtaining a survey map of an unstructured road in a mining area, determining the road boundary of the unstructured road, setting initial nodes in key areas to establish a global node-arc model, and planning an initial global path for vehicle traffic based on the global node-arc model, including: By mapping the mining area, a mapping map of the unstructured roads in the mining area is obtained; Determine the road boundaries and key areas and curve areas of all roads according to the surveying map, and set initial nodes in the key areas and curve areas; All roads included in the surveying and mapping map are used as connecting arcs, the initial node is used as a connecting node, and a node-arc model is constructed; An initial global path for vehicle passage is obtained based on the node-arc model.

3. The method for global and local path hierarchical planning of unstructured roads in mining areas according to claim 1 is characterized in that: The method comprises determining an optimal state of the vehicle during a turning process based on a relationship between the vehicle's posture and the distance from the inner and outer walls of the road when the vehicle is traveling on the curved road section, taking the optimal state of the vehicle during the turning process as a basic condition and the maximum value of the redundant space outside the vehicle as an optimization condition, optimizing the curved road section in the initial local path of the entire road section by ensuring that the redundant space inside and outside the vehicle is the same, and obtaining an optimized local path for the curved road section, including: Acquiring an average curvature of a curved road section and vehicle parameters based on the surveying and mapping map; Based on the average curvature of the curved road section, vehicle parameters, and the relationship between the vehicle's posture and the distance from the inner and outer walls of the road when the vehicle is traveling on the curved road section, the distance L from the outer front end of the vehicle to the outer wall is fixed. 外 , forming different vehicle trajectories and solving different L 外 The shortest distance L between the inner side of the lower vehicle and the inner wall 内 , when the L 外 With L 内 If they are equal, the current vehicle state is determined to be the optimal state of the vehicle during the turning process, and the optimal state of the vehicle is used as the middle position of the optimal path during the turning process; Using the vehicle kinematic model as a constraint, the optimal state of the vehicle as a basic condition, the increment of the front wheel steering angle as a dynamic constraint, and the maximum redundant space outside the vehicle as the optimization condition, the optimal solution is when the redundant space inside and outside the vehicle is consistent. The optimal path for the first half and the optimal path for the second half of the middle position of the curved road section are determined to obtain the optimized local path of the curved road section; The initial node of each curved road section is re-determined on the optimized vehicle trajectory to obtain the starting node and ending node of each curved road section.

4. The method for global and local path hierarchical planning of unstructured roads in mining areas according to claim 3 is characterized in that: The method is based on the difference between the vehicle state and the planned path at the intersection of the trajectories of different road sections, takes the vehicle state in the adjacent road sections as the basic condition, the vehicle kinematic model as the constraint condition, and the planned path difference as zero as the constraint target, adjusts the vehicle posture and the planned trajectory at the intersection of the path trajectories, and optimizes the intersection of the adjacent road sections to obtain the intersection fusion optimized local path, including: Obtain different vehicle states and planned trajectories of adjacent straight and curved road sections at the starting and ending nodes of the curved road sections; Based on the different vehicle states as the basic conditions, the vehicle kinematic model as the constraint condition, and the zero difference between the planned paths at the starting node and the ending node as the constraint target, the vehicle posture and planned trajectory in the straight section are adjusted so that the trajectories at the starting node and the ending node of the curved road section form a smooth transition curve with the planned trajectory of the straight section, thereby obtaining an intersection and fusion optimized local path.

5. The method for global and local path hierarchical planning of unstructured roads in mining areas according to claim 1 is characterized in that: The method of determining a starting position and a target position of the vehicle, determining an optimal global path in the optimized global path according to the starting position and the target position, and determining multiple optimal local paths for the vehicle to follow according to the optimal global path includes: Get the starting position and target position of the vehicle; Determining an optimal global path of a node-arc model from a starting position to a target position in the node-arc model according to a preset global path search algorithm; Determine an optimal global path based on the optimal global path of the node-arc model, and determine multiple local optimal paths based on the optimal global path; The vehicle tracks and drives according to the multiple local optimal paths and the optimal global path.

6. The method for global and local path hierarchical planning of unstructured roads in mining areas according to claim 1, characterized in that: When the vehicle is tracking and driving according to the optimal global path and the multiple optimal local paths, and an obstacle appears on the optimal local path, a temporary local path is re-planned based on the preset vehicle kinematics and artificial potential field method, the vehicle is controlled to track and drive according to the planned temporary local path, and after bypassing the obstacle, the vehicle returns to the optimal global path and the multiple optimal local paths to continue tracking and driving, including: The vehicle tracks and drives from the starting position to the end position according to the preset optimal global path and multiple optimal local paths; When the vehicle sensor senses an obstacle on the optimal local path, determining whether the vehicle can safely pass the obstacle based on road conditions; When the road conditions allow the vehicle to safely pass the obstacle, a temporary local path is replanned based on the preset vehicle kinematics and artificial potential field method. The vehicle tracks and drives along the planned temporary local path, bypasses the obstacle, and returns to the optimal global path and multiple optimal local paths to continue tracking. When the road conditions do not allow the vehicle to safely pass the obstacle, the vehicle returns to the previous node of the current path, takes the previous node as the starting point, and replans the path using the Dijkstra algorithm. The vehicle continues to track according to the obtained temporary optimal global path and temporary multi-segment optimal local paths.

7. The method for global and local path hierarchical planning of unstructured roads in mining areas according to claim 6, characterized in that: The temporary local path planning based on the preset vehicle kinematics and artificial potential field method includes: In the process of temporary local path planning using the artificial potential field method, when the vehicle passes through the obstacle, there is an attraction force of the next node on the vehicle, a repulsion force of the obstacle on the vehicle, and a repulsion force from the boundary of the vehicle. When the combined force of the attraction, repulsion, and boundary repulsion is zero, the vehicle is trapped in a local minimum point and cannot move. Based on the artificial potential field method, an escape force is added to the vehicle to make the vehicle move out of the local minimum point. The direction of the escape force is perpendicular to the attraction of the next node and points to the inside of the road. The escape force gradually increases from zero until the vehicle is driven out of the minimum point, at which point the escape force disappears and the vehicle continues to travel according to the current potential field.

8. The method for hierarchical planning of global and local paths for unstructured roads in mining areas according to claim 7, characterized in that: Determining the vehicle boundary repulsion includes: By establishing a road boundary repulsive potential field, the vehicle is prevented from colliding with the lane boundary during obstacle avoidance. The specific formula for establishing the road boundary repulsive potential field is: in, is the road boundary repulsion, is the proportional gain coefficient, e is a natural constant, s is the shortest distance from the vehicle's center of mass to the global optimal path and the local optimal path, and D is the current lane width when the vehicle is running.

9. A device for global and local path hierarchical planning of unstructured roads in mining areas, characterized in that: include: a map model building module, configured to obtain a survey map of unstructured roads in a mining area, determine the road boundaries of the unstructured roads, set initial nodes in key areas to establish a global node-arc model, and plan an initial global path for vehicle traffic based on the global node-arc model; a path optimization fusion module, configured to split the initial global path into sections according to the initial nodes to obtain initial local paths for the entire section, and optimize and fuse the initial local paths for the entire section based on vehicle kinematics and redundant space on both sides of the vehicle to obtain optimized local paths for the entire section and an optimized global path; A path planning and tracking module is used to determine a starting position and a target position of the vehicle, determine an optimal global path in the optimized global path based on the starting position and the target position, and determine multiple optimal local paths for the vehicle to track based on the optimal global path; a temporary planning and tracking module, configured to, when the vehicle is tracking and driving according to the optimal global path and the multiple optimal local paths, and an obstacle appears on the optimal local path, re-plan the temporary local path based on preset vehicle kinematics and an artificial potential field method, control the vehicle to track and drive according to the planned temporary local path, and return to the optimal global path and the multiple optimal local paths to continue tracking and driving after bypassing the obstacle; The step of splitting the initial global path into sections according to the initial nodes to obtain initial local paths for the entire section, and optimizing and fusing the initial local paths for the entire section based on vehicle kinematics and redundant space on both sides of the vehicle to obtain optimized local paths and an optimized global path for the entire section includes: Splitting the entire road segment of the initial global path into straight road segments and curved road segments according to the initial node to obtain the initial local path of the entire road segment; Based on the vehicle kinematic model and taking the same redundant space between the vehicle and the walls on both sides as the optimization goal, the straight sections in the initial local path of the entire road section are optimized by ensuring that the vehicle has the same redundant space between the walls on both sides, thereby obtaining an optimized local path for the straight sections; determining an optimal state of the vehicle during a turning process based on a relationship between the vehicle's posture and the distance from the inner and outer walls of the road when the vehicle is traveling on the curved road section, taking the optimal state of the vehicle during the turning process as a base condition and the maximum value of the redundant space outside the vehicle as an optimization condition, optimizing the curved section in the initial local path of the entire road section by ensuring that the redundant space inside and outside the vehicle is the same, thereby obtaining an optimized local path for the curved section; Based on the difference between the vehicle state and the planned path at the intersection of the trajectories of different sections, the vehicle state in the adjacent sections is used as the basic condition, the vehicle kinematic model is used as the constraint condition, and the planned path difference is zero as the constraint target. The vehicle posture at the intersection of the path trajectories is adjusted, and the intersection of the adjacent sections is optimized to obtain the intersection fusion optimized local path; The optimized local path for the entire road section and the optimized global path are obtained according to the optimized local path for the straight road section, the optimized local path for the curved road section and the optimized local path for the intersection and fusion.

Images