Multi-target path planning method based on topological map

Abstract

The invention discloses a multi-target path planning method based on a topological map, and relates to the technical field of path planning of automatic driving vehicles, and the method comprises the following steps: S1, constructing a topological map comprising road lengths and important nodes; s2, calculating the distance between the connected road nodes, and establishing an adjacent matrix of the road nodes; s3, selecting a starting point and a plurality of target points; s4, performing multi-target path optimization calculation through an ant colony algorithm, introducing a growth type function, performing adaptive adjustment on ant pheromone volatilization, and designing updating strategies of local pheromones and global pheromones; and S5, designing an evaluation function, and planning an optimal multi-target path by comprehensively considering the conditions of the shortest global path length and the minimum number of optimization iterations of the multi-target ant colony algorithm. By means of the mode, the optimal multi-target path can be found on the topological map through the ant colony algorithm, and the path planning efficiency is improved.

Description

Technical Field

[0001] The present invention relates to a path planning technology for an autonomous driving vehicle, and in particular to a multi-objective path planning method based on a topological map, aiming to improve the efficiency and accuracy of multi-objective path planning. Background Art

[0002] In autonomous driving and intelligent transportation systems, path planning is one of the core issues. Existing technologies mostly focus on single-target path planning, which leads to low computational efficiency in complex environments. Especially in urban roads or multi-target scenarios, existing multi-target path planning methods have problems such as high computational complexity and poor path quality. Therefore, it is of great practical significance to develop an efficient and accurate multi-target path planning method. Summary of the invention

[0003] The present invention provides a multi-objective path planning method based on a topological map, and the specific steps are as follows:

[0004] Step 1: Build a topological map including road length, traffic status and important nodes;

[0005] Step 2: Calculate the distance between connected road nodes and establish the adjacency matrix of road nodes;

[0006] Step 3: Select the starting point and multiple target points;

[0007] Step 4: Use the ant colony algorithm to quickly find the optimal multi-objective path, introduce a growth function, adaptively adjust the ant pheromone volatilization, and design the update strategy of local and global pheromones;

[0008] Step 5: Design a comprehensive evaluation function, taking into account the shortest global path length and the minimum number of multi-objective ant colony algorithm optimization iterations, and plan an optimal multi-objective path.

[0009] Optionally, during the execution of step 1, Amap or other map software can be used to determine the road network structure in a certain area. In the actual abstract topology map, a certain location is used as the starting point of the world map, and other locations are relative locations under the world coordinates. The connectivity between roads is the area where the unmanned vehicle can drive, and the others are obstacle areas or dangerous areas where the unmanned vehicle cannot drive. According to the connectivity of the roads in the map, most of the environmental information is ignored, and only the topological points and edges are retained.

[0010] In the constructed topology map:

[0011] Nodes: represent important locations such as road intersections, destinations, and service stations.

[0012] Edge: represents the road connecting the nodes. Each edge has attributes, including:

[0013] Length: Indicates the actual driving distance of the road.

[0014] Traffic status: reflects the road traffic conditions in real time (such as congestion, closure, etc., which can be obtained through the traffic monitoring system).

[0015] Weight: set according to actual conditions, used for optimization calculation when selecting paths. The cost value represents the distance. The longer the distance, the greater the cost value.

[0016] Optionally, in the method for calculating the distance between connected road nodes in step 2, the distance between road nodes adopts the Euclidean distance formula, which is expressed as follows:

[0017]

[0018] Among them, (x1, y1) is the position coordinate of the first node, and (x2, y2) is the position coordinate of the second connected node.

[0019] Optionally, in the method of establishing the adjacency matrix of road nodes described in step 2, the connectivity of the paths between road nodes can be represented by a matrix, with "0" representing the starting point of the path search, "1" representing the road topology point that can be searched, and "-1" representing the road topology point that cannot be searched, indicating that there is no connectivity between the search starting point and the topology point.

[0020] Optionally, in step 4, in the process of performing multi-objective path fast optimization calculation by using the ant colony algorithm, the steps specifically include:

[0021] Step 4.1: Initialize the parameters of the ant colony algorithm, including the number of ants N, the initial pheromone concentration τ0, the pheromone volatility ρ, the heuristic function weight factors α and β, and the maximum number of iterations T max .

[0022] Step 4.2: Release the ant, find the target point, and select the next node by the following probability:

[0023]

[0024] Among them, η ij Represents the heuristic value from node i to node j.

[0025] Optionally, the growth function introduced in step 4 is used to adaptively adjust the volatilization amount of ant pheromone, and the growth function expression is:

[0026] g(t)=γ·f(t)

[0027] Among them, γ is the growth coefficient, and f(t) is the time function, which reflects the quality of path selection and the dynamic changes of pheromone volatilization.

[0028] The pheromone evaporation rate is adaptively adjusted according to the path quality, so that the nodes with better path selection can obtain more pheromone support in subsequent iterations. In the early stage of algorithm convergence, the pheromone evaporation coefficient obtained by the growth function is small, and the guiding effect of the ant colony on each ant is weak, thereby increasing the global solution space searched by the algorithm and improving the search accuracy; in the later stage of algorithm convergence, the pheromone evaporation coefficient obtained by the growth function increases, the guiding effect of the ant colony on each ant is strengthened, and the convergence speed of the ant colony searching for the optimal global path is significantly accelerated.

[0029] Furthermore, the local pheromone update strategy in step 4 is characterized in that after each ant completes path selection, the local pheromone is updated according to the path length L. The specific update formula is:

[0030] τ ij =(1-ρ)·τ ij +Δτ ij

[0031] in, Q is a constant, ρ is the information volatilization rate, and the shorter the path, the more pheromone is updated.

[0032] Furthermore, the global pheromone update strategy in step 4 is specifically to update the global pheromone according to the path quality of all ants after all ants complete path selection. The specific update formula is:

[0033]

[0034] in, is the pheromone increment generated when the kth ant passes through the path node (i, j). Ensure that when the global pheromone is updated, the higher the path quality, the more pheromone is increased, which promotes the spread of high-quality paths.

[0035] Optionally, in the method of designing the evaluation function in step 5, comprehensively considering the shortest global path length and the number of optimization iterations of the multi-objective ant colony algorithm, the evaluation function expression is:

[0036]

[0037] Where L is the path length, n is the number of target points passed, N is the total number of target points, Τ is the number of algorithm iterations, and ω1, ω2, ω3 are weight coefficients.

[0038] After multiple rounds of iterations, the optimal solution is gradually converged by adjusting the pheromone concentration and the path selection probability.

[0039] The present invention provides a multi-objective path planning method based on a topological map, which relates to the technical field of path planning for autonomous driving vehicles. First, a topological map including road lengths and important nodes is constructed; the distances between connected road nodes are calculated, and an adjacency matrix of road nodes is established; a starting point and multiple target points are selected; a multi-objective path fast-travel optimization calculation is performed through an ant colony algorithm, a growth function is introduced, the volatilization of ant pheromones is adaptively adjusted, and local pheromone and global pheromone update strategies are designed; an evaluation function is designed, and the shortest global path length and the minimum number of optimization iterations of the multi-objective ant colony algorithm are comprehensively considered as conditions to plan the optimal multi-objective path. In the above manner, the present invention can find an optimal multi-objective path through an ant colony algorithm on a topological map, thereby improving the efficiency of algorithm planning. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0041] Figure 1 It is a schematic flow chart of a multi-objective path planning method based on a topological map of the present invention.

[0042] Figure 2 It is a schematic diagram of the topology map of the present invention.

[0043] Figure 3 It is a partial adjacency matrix schematic diagram of the present invention.

[0044] Figure 4 It is a schematic diagram of the multi-objective path planning result of the present invention.

[0045] Figure 5 It is a schematic diagram of multi-objective path planning simulation of the present invention. DETAILED DESCRIPTION

[0046] The following is a detailed description of implementation cases of the present invention, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0047] The following are some nouns and terminology descriptions of the present invention:

[0048] ACO: Ant Colony Optimization, ant colony algorithm.

[0049] See also Figure 1 The present invention provides a multi-objective path planning method based on a topological map, comprising the following steps:

[0050] Step 1: Build a topological map including road length, traffic status and important nodes;

[0051] Step 2: Calculate the distance between connected road nodes and establish the adjacency matrix of road nodes;

[0052] Step 3: Select the starting point and multiple target points;

[0053] Step 4: Use the ant colony algorithm to quickly find the optimal multi-objective path, introduce a growth function, adaptively adjust the ant pheromone volatilization, and design the update strategy of local and global pheromones;

[0054] Step 5: Design a comprehensive evaluation function, taking into account the shortest global path length and the minimum number of multi-objective ant colony algorithm optimization iterations, and plan the optimal multi-objective path.

[0055] The following is a further explanation based on the specific implementation steps:

[0056] Step 1: Build a topology map.

[0057] Use Amap or other map software to determine the road network structure in a certain area. In the actual abstract topology map, a certain location is the starting point of the world map, and other locations are relative locations under the world coordinates. The connectivity between the roads is the area where the unmanned vehicle can drive, and the others are obstacle areas or dangerous areas where the unmanned vehicle cannot drive. According to the connectivity of the roads in the map, most of the environmental information is ignored, and only the topological points and edges are retained. For details, please refer to Figure 2 Schematic diagram of the topological map provided.

[0058] In the constructed topology map:

[0059] Nodes: represent important locations such as road intersections, destinations, and service stations.

[0060] Edge: represents the road connecting the nodes. Each edge has attributes, including:

[0061] Length: Indicates the actual driving distance of the road.

[0062] Traffic status: reflects the road traffic conditions in real time (such as congestion, closure, etc., which can be obtained through the traffic monitoring system).

[0063] Weight: It is set according to the actual situation and is used for optimization calculation when selecting a path. The cost value represents the distance. The longer the distance, the greater the cost value.

[0064] Step 2: Calculate the distances between connected road nodes and establish the adjacency matrix of road nodes.

[0065] The distance between road nodes adopts the Euclidean distance formula, which is expressed as:

[0066]

[0067] Among them, (x1, y1) is the position coordinate of the first node, and (x2, y2) is the position coordinate of the second connected node.

[0068] The connectivity of the paths between road nodes can be represented by a matrix, with "0" representing the path search starting point, "1" representing the road topology point that can be searched, and "-1" representing the road topology point that cannot be searched, indicating that there is no connectivity between the search starting point and the topology point. For details, please refer to Figure 3 A partial diagram of the adjacency matrix provided.

[0069] Step 3: Select the starting point and the destination point.

[0070] By inputting the starting point and multiple target points through the user interaction module, the system automatically selects important destinations to ensure the effectiveness of route planning.

[0071] Step 4: The ant colony algorithm performs fast multi-target path optimization calculations, passing through each target point only once to find the shortest path. The specific steps include:

[0072] Step 4.1: Initialize the parameters of the ant colony algorithm:

[0073] The number of ants N (for example, 20 ants);

[0074] initial pheromone concentration τ0 (e.g., 0.1);

[0075] pheromone volatility ρ (e.g., 0.1);

[0076] Heuristic function weight factors α and β (e.g., 0.1);

[0077] Maximum number of iterations T max (For example, 0.1).

[0078] Step 4.2: Release the ant and select the next node by the following probability:

[0079]

[0080] Among them, η ij Represents the heuristic value from node i to node j.

[0081] Record the path and length taken by the ants.

[0082] Step 4.3: Adaptively adjust ant pheromone volatilization.

[0083] The introduced growth function is used to adaptively adjust the volatilization of ant pheromone. The growth function expression is:

[0084] g(t)=γ·f(t)

[0085] Among them, γ is the growth coefficient, and f(t) is the time function, which reflects the quality of path selection and the dynamic changes of pheromone volatilization.

[0086] The pheromone evaporation rate is adaptively adjusted according to the path quality, so that the nodes with better path selection can obtain more pheromone support in subsequent iterations. In the early stage of algorithm convergence, the pheromone evaporation coefficient obtained by the growth function is small, and the guiding effect of the ant colony on each ant is weak, thereby increasing the global solution space searched by the algorithm and improving the search accuracy; in the later stage of algorithm convergence, the pheromone evaporation coefficient obtained by the growth function increases, the guiding effect of the ant colony on each ant is strengthened, and the convergence speed of the ant colony searching for the optimal global path is significantly accelerated.

[0087] Step 4.4: Local pheromone update

[0088] After each ant completes path selection, it updates the local pheromone according to the path length L. The specific update formula is:

[0089] τ ij =(1-ρ)·τ ij +Δτ ij

[0090] in, Q is a constant, ρ is the information volatilization rate, and the shorter the path, the more pheromone is updated.

[0091] Step 4.5: Global pheromone update.

[0092] After all ants have completed path selection, the global pheromone update is performed based on the path quality of all ants. The specific update formula is:

[0093]

[0094] in, is the pheromone increment generated when the kth ant passes through the path node (i, j). Ensure that when the global pheromone is updated, the higher the path quality, the more pheromone is increased, which promotes the spread of high-quality paths.

[0095] Step 5: Calculate the evaluation function, taking into account the shortest global path length and the number of iterations of the multi-objective ant colony algorithm optimization. For details, please refer to Figure 4The schematic diagram of the multi-objective path planning results provided, in which S is the starting point, and G1, G2, G3, G4, and G5 are multiple target points; Figure 5 Schematic diagram of multi-objective path planning simulation.

[0096] The evaluation function expression is:

[0097]

[0098] Where L is the path length, n is the number of target points passed, N is the total number of target points, Τ is the number of algorithm iterations, and ω1, ω2, ω3 are weight coefficients.

[0099] After multiple rounds of iterations, by adjusting the pheromone concentration and the probability of path selection, it gradually converges to the optimal solution. In this way, the ant colony algorithm can be used to find an optimal multi-objective path on the topological map, which improves the efficiency of algorithm planning.

[0100] What is disclosed above is only a preferred embodiment of the present invention, and it certainly cannot be used to limit the scope of rights of the present invention. A person skilled in the art can understand that all or part of the processes of the above embodiment and equivalent changes made according to the claims of the present invention still fall within the scope of the invention.

Claims

1. A multi-objective path planning method based on topological map, characterized in that: The following steps are involved: Step 1: Build a topological map including road length, traffic status and important nodes; Step 2: Calculate the distance between connected road nodes and establish the adjacency matrix of road nodes; Step 3: Select the starting point and multiple target points; Step 4: Use the ant colony algorithm to quickly find the optimal multi-objective path, introduce a growth function, adaptively adjust the ant pheromone volatilization, and design the update strategy of local and global pheromones; Step 5: Design an evaluation function, comprehensively consider the shortest global path length and the minimum number of multi-objective ant colony algorithm optimization iterations, and plan the optimal multi-objective path.

2. The method according to claim 1, characterized in that In the process of constructing the topological map described in step 1, Amap or other map software can be used to determine the road traffic network structure in a certain area. In the actual abstract topological map, a certain location is used as the starting point of the world map, and other locations are relative positions under the world coordinates. The connectivity between roads is the area where unmanned vehicles can drive, and the others are obstacle areas or dangerous areas where unmanned vehicles are not feasible. According to the connectivity of the roads in the map, most of the environmental information is ignored, and only the topological points and edges are retained.

3. The method according to claim 2, characterized in that In the above-mentioned topological map construction, important roads, nodes, important buildings and the length of segmented roads are represented by nodes and edges. Nodes are road intersections, and edges are roads connecting nodes. The attributes of edges include information such as length, traffic status and weight. The edge length in the topological map represents the path, and the cost value represents the distance. The longer the distance, the greater the cost value.

4. The method according to claim 1, characterized in that In the method for calculating the distance between connected road nodes described in step 2, the distance between road nodes adopts the Euclidean distance formula, which is expressed as follows: Among them, (x1, y1) is the position coordinate of the first node, and (x2, y2) is the position coordinate of the second connected node.

5. The method according to claim 1, characterized in that: In the method of establishing the adjacency matrix of road nodes described in step 2, the connectivity of the paths between road nodes can be represented by a matrix, with "0" representing the starting point of the path search, "1" representing the road topology point that can be searched, and "-1" representing the road topology point that cannot be searched, indicating that there is no connectivity between the search starting point and the topology point.

6. The method according to claim 1, characterized in that: In step 4, the ant colony algorithm is used to perform a multi-objective path fast optimization calculation process, and the steps specifically include: Step 1: Initialize the parameters of the ant colony algorithm, including the number of ants N, the initial pheromone concentration τ0, the pheromone volatility ρ, the heuristic function weight factors α and β, and the maximum number of iterations T max . Step 2: Release the ant, find the target point, and select the next node by the following probability: Among them, η ij Represents the heuristic value from node i to node j.

7. The method according to claim 1, characterized in that The growth function introduced in step 4 is used to adaptively adjust the volatilization amount of ant pheromone. The growth function expression is: g(t)=γ·f(t) Among them, γ is the growth coefficient, and f(t) is the time function, which reflects the quality of path selection and the dynamic changes of pheromone volatilization. The pheromone evaporation rate is adaptively adjusted according to the path quality, so that the nodes with better path selection can obtain more pheromone support in subsequent iterations. In the early stage of algorithm convergence, the pheromone evaporation coefficient obtained by the growth function is small, and the guiding effect of the ant colony on each ant is weak, thereby increasing the global solution space searched by the algorithm and improving the search accuracy; in the later stage of algorithm convergence, the pheromone evaporation coefficient obtained by the growth function increases, the guiding effect of the ant colony on each ant is strengthened, and the convergence speed of the ant colony searching for the optimal global path is significantly accelerated.

8. The method according to claim 1, characterized in that The local pheromone update strategy in step 4 is characterized in that after each ant completes path selection, the local pheromone is updated according to the path length L. The specific update formula is: t ij =(1-ρ)·τ ij +Δt ij in, Q is a constant, ρ is the information volatilization rate, and the shorter the path, the more pheromone is updated.

9. The method according to claim 1, characterized in that: The global pheromone update strategy described in step 4 is to perform global pheromone update according to the path quality of all ants after all ants complete path selection. The specific update formula is: in, is the pheromone increment generated when the kth ant passes through the path node (i, j). Ensure that when the global pheromone is updated, the higher the path quality, the more pheromone is increased, which promotes the spread of high-quality paths.

10. The method according to claim 1, characterized in that In the method of designing the evaluation function in step 5, which comprehensively considers the shortest global path length and the number of optimization iterations of the multi-objective ant colony algorithm, the evaluation function expression is: Where L is the path length, n is the number of target points passed, N is the total number of target points, Τ is the number of algorithm iterations, and ω1, ω2, ω3 are weight coefficients. After multiple rounds of iterations, the optimal solution is gradually converged by adjusting the pheromone concentration and the path selection probability.

Images