Unmanned mine car loading path planning method

Abstract

The invention relates to the field of automatic driving, in particular to an unmanned mine car loading path planning method, which comprises the following steps of: 1, acquiring nodes, obstacle information, an initial position, a current position, an orientation and a target position of a map; 2, generating a plurality of local nodes based on an A * search algorithm; step 3, determining a standard node from the plurality of local nodes according to the current position; step 4, generating a plurality of speed commands conforming to physical constraints of the unmanned mine car based on the current position and orientation of the current unmanned mine car, determining an optimal speed command according to a predefined path evaluation function, executing the optimal speed command, updating the current position and orientation information of the unmanned mine car, judging whether the current position reaches a target node, and if the current position reaches the target node, executing the optimal speed command; if yes, updating the target node according to the current position; and 5, repeatedly executing the step 4 until the target position is reached. According to the invention, a more accurate and highly adaptive driving path is effectively generated, and the transportation efficiency and safety of the unmanned mine car are improved.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving, and in particular to a method for planning a loading path for an unmanned mining vehicle. Background Art

[0002] Unmanned mining vehicles are mining vehicles that use automation and unmanned driving technology and are used to transport ore or soil in mines or other excavation areas. Path planning and obstacle avoidance of unmanned mining vehicles are key links in unmanned mining vehicle technology, with the goal of enabling the mining vehicle to automatically travel from the starting point to the target point while avoiding obstacles on the way. The purpose of path planning and obstacle avoidance is to improve the transportation efficiency of mining vehicles, reduce transportation costs, ensure transportation safety, and reduce dependence on human resources and improve work efficiency through automated control.

[0003] Existing unmanned minecart path planning methods usually use some classic path search algorithms, such as the Dijkstra algorithm. These path search algorithms search for a path from the starting point to the target point in a given map based on the starting point and target point of the unmanned minecart. However, the main disadvantage of these methods is that they usually only consider static environmental information, such as the location of the map and nodes, but do not consider the dynamic information of the unmanned minecart, such as speed, angular velocity, etc. Therefore, the paths planned by these methods may be inaccurate in actual execution and cannot adapt to complex and dynamically changing environments. Summary of the invention

[0004] In order to solve the above problems, the present invention provides a method for planning a loading path for an unmanned mining vehicle.

[0005] The unmanned mine vehicle loading path planning method comprises:

[0006] Step 1: Get the map's nodes, obstacle information, starting position, current position, direction, and target position;

[0007] Step 2: Generate multiple local nodes based on the A* search algorithm according to the map nodes, obstacle information, starting position and target position;

[0008] Step 3: According to the current position, determine the next local node from multiple local nodes as the target node;

[0009] Step 4: Based on the current position and orientation of the unmanned mine car, generate multiple speed commands that meet the physical constraints of the unmanned mine car, determine the optimal speed command and the predicted optimal position and predicted orientation corresponding to the optimal speed command from the multiple speed commands according to the predefined path evaluation function, execute the optimal speed command, update the current position and orientation information of the unmanned mine car to the predicted optimal position and the predicted orientation respectively, judge whether the current position reaches the target node, and if so, update the target node according to the current position;

[0010] Step 5: Repeat step 4 until the target position is reached.

[0011] Furthermore, step 2 includes:

[0012] Step 2A, define the heuristic function of the A* search algorithm;

[0013] Step 2B, using the A* search algorithm, based on the nodes, starting positions and target positions of the map, multiple global nodes are generated, and a global node set is formed in the order of the path, wherein the first global node in the global node set is the head global node, and the last global node is the target global node;

[0014] Step 2C: Filter the global node set according to the obstacle information to obtain multiple local nodes.

[0015] Furthermore, the specific steps of step 2A include:

[0016] The heuristic function of the A* search algorithm is defined as:

[0017] f(n)=g(n)+a(x)[h(n)+h(f)];

[0018]

[0019] μ = Dist(S, G);

[0020] x=h(n);

[0021] Among them, f(n) represents the cost of the current node n, h(n) represents the estimated distance from the current node to the target position, g(n) represents the distance from the starting position to the current node n, x represents the search node, a(x) represents the function of dynamically adjusting the weight, h(f) represents the estimated distance from the parent node f of the current node to the target position, Dist(S,G) represents the distance from the starting position to the target position, and σ and μ are intermediate variables.

[0022] Furthermore, the specific steps of step 2C include:

[0023] Traverse each global node in the global node set except the first global node and the target global node as the calculation node m;

[0024] Calculate the area of ​​the triangle enclosed by the computational node m and its two adjacent global nodes;

[0025] Determine the area of ​​a triangle;

[0026] If the area is smaller than a preset first threshold, the computing node m is deleted from the global node set.

[0027] Furthermore, the specific steps of step 2C include:

[0028] Traverse each global node in the global node set except the first global node and the target global node as the calculation node m;

[0029] Connect two global nodes adjacent to the computing node m;

[0030] Determine whether its connection line passes through the obstacle area. If not, delete the computing node m from the global node set.

[0031] Furthermore, the path evaluation function is:

[0032]

[0033] Among them, v is the speed of the unmanned mine car, w is the angular velocity of the unmanned mine car, H(v,w) is the heading angle evaluation subfunction, α is the heading angle evaluation weight, S(v,w) is the speed evaluation subfunction, β is the speed evaluation weight, D(v,w) is the obstacle distance evaluation subfunction, γ is the obstacle distance evaluation weight, G(v,w) is the global path evaluation subfunction, η is the global path evaluation weight, P is the position coordinate of the unmanned mine car predicted by v and w at time t, f(P) is the global path evaluation subfunction, g(P) represents the distance from the current position to the starting position; h(P) represents the distance from the current position to the target node.

[0034] Furthermore, the obstacle distance evaluation subfunction D(v,w) is:

[0035]

[0036] Among them, Dist(P,O) is the minimum distance between the unmanned mine car and the obstacle, and R represents the safe distance for the unmanned mine car to travel.

[0037] Furthermore, when the value of the obstacle distance evaluation subfunction D(v,w) does not exceed the unmanned vehicle's safe braking distance S_Dist(v), let D(v,w)=0, where S_Dist(v) is:

[0038]

[0039] Among them, a s is the maximum braking acceleration of the unmanned mining vehicle, T s is the time required for the unmanned vehicle to slow down from its travel speed to a complete standstill, and t is the braking acceleration time of the unmanned vehicle.

[0040] Furthermore, the heading angle evaluation weight α exceeds the obstacle distance evaluation weight γ under any of the following conditions:

[0041] When the distance between the obstacle and the current position of the unmanned minecart exceeds 1.5 times the minimum turning radius of the unmanned minecart;

[0042] Alternatively, a line connecting the current position of the unmanned mining vehicle and the predicted optimal position does not cross an obstacle area;

[0043] Alternatively, the distance between the current position of the unmanned mine car and the predicted optimal position does not exceed 0.5 times the minimum turning radius of the unmanned mine car.

[0044] One or more technical solutions provided in the embodiments of the present invention have at least the following technical effects or advantages:

[0045] The present invention combines global path planning and local path planning, fully considers the dynamic state of the unmanned vehicle and the static information of the environment, effectively generates a more accurate and adaptable driving path, and improves the transportation efficiency and safety of the unmanned mining vehicle. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 A flow chart of a method for planning a loading path for an unmanned mine vehicle provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0047] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. Before describing in detail the technical solutions of each embodiment of the present invention, the nouns and terms involved are explained. In this specification, components with the same name or the same number represent similar or identical structures and are only for illustrative purposes.

[0048] The method of the present invention is as follows Figure 1 As shown, the specific steps are as follows:

[0049] 1. Data Acquisition

[0050] When optimizing the loading path planning, the unmanned mining truck first needs to obtain the corresponding map information through the cloud control platform. As the soul hub of unmanned driving, the cloud control platform integrates resources such as vehicles, roadside infrastructure, and communication networks, providing a series of powerful support and services for unmanned mining trucks. In this process, the cloud control platform can not only monitor the operating status of the unmanned mining truck in real time, but also collect and analyze traffic data, providing the necessary data support and computing power for the unmanned mining truck.

[0051] Based on the cloud control platform, the unmanned mining vehicle can obtain map information of the open-pit mine loading area, including map nodes, obstacle information, starting position, current position, direction and target position, etc. At the same time, the unmanned mining vehicle can also obtain real-time speed and angular velocity information with the help of the vehicle-road V2X sensor, and interact with the cloud control platform in real time to facilitate more optimized path planning.

[0052] 2. Global path planning

[0053] The main purpose of global path planning is to determine the approximate path of the unmanned minecart from the starting position to the target position. This path needs to take into account obstacles in the map, as well as the positions of the starting position and the target position. Global path planning provides a feasible path from the starting position to the target position, but this path may not be optimal because it does not take into account dynamic factors such as the speed and angular velocity of the unmanned minecart. The results of global path planning will be used as the basis for subsequent local path planning.

[0054] The present invention optimizes the A* search algorithm to achieve global path planning. The A* search algorithm is a heuristic search algorithm that selects the optimal search direction by estimating the minimum cost from the starting position to the target position. The optimized version of the A* algorithm of the present invention dynamically adjusts the heuristic function according to the distance from the node to the target position to speed up the search. Specifically, if the node is far from the target position, the heuristic function will increase, making the search more focused; conversely, if the node is close to the target position, the heuristic function will decrease, expanding the search range.

[0055] The heuristic function of the optimized version of the A* algorithm of the present invention is:

[0056] f(n)=g(n)+a(x)[h(n)+h(f)];

[0057]

[0058] μ = Dist(S, G);

[0059] x = h(n);

[0060] Among them, f(n) represents the cost value of the current node n, h(n) represents the estimated distance from the current node to the target position, g(n) represents the distance from the starting position to the current node n, x represents the search node, a(x) represents the function of dynamically adjusting the weight, h(f) represents the estimated distance from the parent node f of the current node to the target position, Dist(S,G) represents the distance from the starting position to the target position, σ and μ are intermediate variables. During the search process, each node has a parent node, which is the node that directly precedes the current node in the search path.

[0061] Under the guidance of the heuristic function, the global path planning finally calculates a set of multiple sorted global nodes from the first global node to the target global node, that is, the global node set, where the first global node corresponds to the starting position, the target global node corresponds to the target position, and the sorting order in the global node set is from small to large according to the distance from the first global node.

[0062] 3. Local node calculation

[0063] After the global path planning is completed, key points need to be extracted on this global path, which will be used as local nodes.

[0064] 3.1 Area extraction method

[0065] If the area of ​​the triangle enclosed by three adjacent global nodes in the path is approximately 0, then the three global nodes are actually in a straight line, and the middle global node can be regarded as a redundant global node.

[0066] Starting from the first global node, select the global node closest to the first global node as the computing node m, calculate the area of ​​the triangle enclosed by the computing node m and its two adjacent global nodes (global node m-1 and global node m+1), and if the area is less than the preset first threshold, delete the computing node m from the global node set. Move the node m one position toward the target global node in turn, and repeat the above steps until m is an adjacent global node of the target global node. In this embodiment, the first threshold is 0.5.

[0067] 3.2 Screening key points

[0068] If the straight path between non-adjacent global nodes does not pass through the obstacle area, then the global nodes between them can be regarded as redundant global nodes.

[0069] Starting from the first global node, select the global node closest to the first global node as the calculation node m, connect its two adjacent global nodes (global node m-1 and global node m+1), and determine whether the connection line passes through the obstacle area. If not, delete the calculation node m from the global node set. Move node m one position toward the target global node in turn, and repeat the above steps until m is the adjacent global node of the target global node.

[0070] All remaining global nodes are regarded as all local nodes, and the first global node and the target global node are both local nodes.

[0071] 4. Local path planning

[0072] According to the current position of the unmanned mine car, its next local node is determined as the target node. In each time period, the Dynamic Window Approach algorithm (DWA algorithm) is used for local path planning based on the current position and orientation of the unmanned mine car and the position of the target node.

[0073] Time-step path planning: Based on the current position and orientation of the current unmanned mine car, a set of possible speed commands are generated, which meet the physical constraints of the unmanned mine car, such as maximum speed, maximum steering angle, etc. For each speed command, its calculated path within a preset time period is predicted, and the cost value of each calculated path is calculated according to the preset path evaluation function. Among all the calculated paths, the speed command corresponding to the calculated path with the smallest cost value is selected as the optimal speed command, and the predicted optimal position and predicted orientation corresponding to the optimal speed command are calculated. The optimal speed command is executed, and the current position and orientation information of the unmanned mine car are updated to the predicted optimal position and the predicted orientation respectively.

[0074] Determine whether the unmanned mining vehicle has reached the target node. If it has reached the target node, determine whether it has reached the target global node. If it has reached it, end the path planning process. If it has not reached it, find the next local node of the target node and set it as the target node, and continue to execute the time-step path planning steps; if it has not reached the target node, repeat the time-step path planning steps.

[0075] The path evaluation function is:

[0076]

[0077] Among them, v is the speed of the unmanned mine car, w is the angular velocity of the unmanned mine car, H(v,w) is the heading angle evaluation subfunction, α is the heading angle evaluation weight, S(v,w) is the speed evaluation subfunction, β is the speed evaluation weight, D(v,w) is the obstacle distance evaluation subfunction, γ is the obstacle distance evaluation weight, G(v,w) is the global path evaluation subfunction, η is the global path evaluation weight, P is the position coordinate of the unmanned mine car predicted by v and w at time t, f(P) is the global path evaluation subfunction, g(P) represents the distance from the current position to the starting position; h(P) represents the distance from the current position to the target node.

[0078] The obstacle distance evaluation subfunction D(v,w) is:

[0079]

[0080] Where Dist(P,O) is the minimum distance between the unmanned mine car and the obstacle, and R represents the safe distance for the unmanned mine car to travel. If D(v,w)≤S_Dist(v), then let D(v,w)=0, where S_Dist(v) represents the safe braking distance of the unmanned mine car:

[0081]

[0082] Among them, a s is the maximum braking acceleration of the unmanned mining vehicle, T s is the time required for the unmanned vehicle to slow down from its travel speed to a complete standstill, and t is the braking acceleration time of the unmanned vehicle.

[0083] In order to solve the problem of long calculation time caused by lengthy planning routes, the adaptive strategy of the weight of the evaluation sub-function is adjusted. When the environment of the unmanned mine car meets any of the following three conditions, let α>γ:

[0084] 1) When obstacles i The distance from the current position P of the unmanned mining vehicle satisfies Dist(P,ob i )>1.5L, it means that there are no obstacles in the mining area that is 1.5 times the minimum turning radius L of the unmanned mining vehicle;

[0085] 2) When the current position of the unmanned vehicle P(x P ,y P ) and the target point G(x G ,y G ) between obstacles (ob i (x obi ,y obi ))∈{(x,y)|x P ≤x≤x G ,y P ≤y≤y G), the line connecting the unmanned mining vehicle and the predicted optimal position does not cross the obstacle area;

[0086] 3) When Dist(P,G)≤0.5L, it means that the unmanned mining vehicle is very close to the predicted optimal position.

[0087] The time-step path planning steps are repeated until the target position is reached.

[0088] The above-described embodiments are merely descriptions of preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. A method for planning a loading path for an unmanned mining vehicle, characterized in that: The following steps are involved: Step 1: Get the map's nodes, obstacle information, starting position, current position, direction, and target position; Step 2: Generate multiple local nodes based on the A* search algorithm according to the map nodes, obstacle information, starting position and target position; Step 3: According to the current position, determine the next local node from multiple local nodes as the target node; Step 4: Based on the current position and orientation of the unmanned mine car, generate multiple speed commands that meet the physical constraints of the unmanned mine car, determine the optimal speed command and the predicted optimal position and predicted orientation corresponding to the optimal speed command from the multiple speed commands according to the predefined path evaluation function, execute the optimal speed command, update the current position and orientation information of the unmanned mine car to the predicted optimal position and the predicted orientation respectively, judge whether the current position reaches the target node, and if so, update the target node according to the current position; Step 5: Repeat step 4 until the target position is reached.

2. The method for planning the loading path of an unmanned mining vehicle according to claim 1, characterized in that: Step 2 includes: Step 2A, define the heuristic function of the A* search algorithm; Step 2B, using the A* search algorithm, based on the nodes, starting positions and target positions of the map, multiple global nodes are generated, and a global node set is formed in the order of the path, wherein the first global node in the global node set is the head global node, and the last global node is the target global node; Step 2C: Filter the global node set according to the obstacle information to obtain multiple local nodes.

3. The method for planning the loading path of an unmanned mine vehicle according to claim 2, characterized in that: The specific steps of Step 2A include: The heuristic function of the A* search algorithm is defined as: f(n)=g(n)+a(x)[h(n)+h(f)]; μ = Dist(S, G); x = h(n); Among them, f(n) represents the cost of the current node n, h(n) represents the estimated distance from the current node to the target position, g(n) represents the distance from the starting position to the current node n, x represents the search node, a(x) represents the function of dynamically adjusting the weight, h(f) represents the estimated distance from the parent node f of the current node to the target position, Dist(S,G) represents the distance from the starting position to the target position, and σ and μ are intermediate variables.

4. The method for planning the loading path of an unmanned mine vehicle according to claim 2, characterized in that: The specific steps of step 2C include: Traverse each global node in the global node set except the first global node and the target global node as the calculation node m; Calculate the area of ​​the triangle enclosed by the computational node m and its two adjacent global nodes; Determine the area of ​​a triangle; If the area is smaller than a preset first threshold, the computing node m is deleted from the global node set.

5. The method for planning the loading path of an unmanned mine vehicle according to claim 2, characterized in that: The specific steps of step 2C include: Traverse each global node in the global node set except the first global node and the target global node as the calculation node m; Connect two global nodes adjacent to the computing node m; Determine whether its connection line passes through the obstacle area. If not, delete the computing node m from the global node set.

6. The unmanned mine vehicle loading path planning method according to claim 1 is characterized in that: The path evaluation function is: Among them, v is the speed of the unmanned mine car, w is the angular velocity of the unmanned mine car, H(v,w) is the heading angle evaluation subfunction, α is the heading angle evaluation weight, S(v,w) is the speed evaluation subfunction, β is the speed evaluation weight, D(v,w) is the obstacle distance evaluation subfunction, γ is the obstacle distance evaluation weight, G(v,w) is the global path evaluation subfunction, η is the global path evaluation weight, P is the position coordinate of the unmanned mine car predicted by v and w at time t, f(P) is the global path evaluation subfunction, g(P) represents the distance from the current position to the starting position; h(P) represents the distance from the current position to the target node.

7. The unmanned mine vehicle loading path planning method according to claim 6 is characterized in that: The obstacle distance evaluation subfunction D(v,w) is: Among them, Dist(P,O) is the minimum distance between the unmanned mine car and the obstacle, and R represents the safe distance for the unmanned mine car to travel.

8. The method for planning the loading path of an unmanned mining vehicle according to claim 6, characterized in that: When the value of the obstacle distance evaluation subfunction D(v,w) does not exceed the unmanned vehicle's safe braking distance S_Dist(v), let D(v,w)=0, where S_Dist(v) is: Among them, a s is the maximum braking acceleration of the unmanned mining vehicle, T s is the time required for the unmanned vehicle to slow down from its travel speed to a complete standstill, and t is the braking acceleration time of the unmanned vehicle.

9. The unmanned mine vehicle loading path planning method according to claim 6, characterized in that: The heading angle evaluation weight α exceeds the obstacle distance evaluation weight γ under any of the following conditions: When the distance between the obstacle and the current position of the unmanned minecart exceeds 1.5 times the minimum turning radius of the unmanned minecart; Alternatively, a line connecting the current position of the unmanned mining vehicle and the predicted optimal position does not cross an obstacle area; Alternatively, the distance between the current position of the unmanned mine car and the predicted optimal position does not exceed 0.5 times the minimum turning radius of the unmanned mine car.

Images