Test path planning method, automatic driving road test method, device and related equipment
By splitting the edges of interest in the road network diagram and combining the shortest path search algorithm, the road test route of autonomous driving is generated and optimized, and the problem of time-consuming and labor-consuming route design in the existing technology is solved, and the focus is achieved on high-incidence scenarios of accidents Improvement of testing and testing efficiency.
Patent Information
- Application Number
- CN202210857794.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-07-20
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2042-07-20
AI Technical Summary
The existing autonomous driving road test route design relies on manual experience, making it difficult to optimize routes to cover high-incidence scenarios, and it is time-consuming and labor-intensive, making it difficult to achieve the optimal test results.
By splitting the interested edges in the road network diagram into multiple extended edges, combining the starting point, end point and passing point, using the shortest path search algorithm for path planning, generating an extended test route, and replacing it with the interested edge with heavy Create a road network map to realize key tests on the road sections of interest.
The key test of the section of road of interest is realized, the length of the test route is shortened, the testing efficiency is improved, the complexity of the manual design route is reduced, and most pathfinding algorithms are adapted to facilitate subsequent improvements.
Smart Images

Figure CN115218917B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of autonomous driving technology, and more specifically, to a test path planning method, an autonomous driving road test method, a device and related equipment. Background Art
[0002] In the development of autonomous driving technology, putting autonomous driving vehicles on actual roads for testing (road testing) is an important part of algorithm optimization. By sending vehicles to test on designated roads, the availability and accuracy of the algorithm can be verified, and the algorithm can be optimized based on the data from the road test. Therefore, it is very important to specify a road test route that covers as many accident-prone scenarios as possible for the original algorithm.
[0003] Generally, for current autonomous driving road test routes, it is necessary to manually specify the route based on experience or areas of interest, and then specify the vehicle. This processing method is relatively labor-intensive, and it is often difficult to design the optimal test route to cover the sections of interest as much as possible and to use the test time efficiently, making it difficult to achieve the best test results. Summary of the invention
[0004] In view of this, the present application provides a test path planning method, an autonomous driving road test method, an apparatus and related equipment to realize the automation of path planning and autonomous driving road testing.
[0005] To achieve the above objectives, the present application provides a test path planning method in a first aspect, comprising:
[0006] Splitting each edge of interest in the road network graph into a plurality of extended edges connected in series, to obtain an extended road network graph, wherein the total length of the plurality of extended edges is equal to the length of the edge of interest;
[0007] Based on the extended road network graph, in combination with the starting point, the end point and the set of waypoints, a preset shortest path search algorithm is used to perform path planning to obtain an extended test route corresponding to the extended road network graph;
[0008] Each extended edge in the extended test route is replaced with an edge of interest corresponding to the extended edge, so as to obtain a test route corresponding to the road network graph.
[0009] Preferably, the interest level of the edge of interest is determined by the weight of the edge, and the process of splitting each edge of interest in the road network graph into a plurality of extended edges connected in series to obtain a reconstructed road network graph includes:
[0010] Based on the weight of each edge of interest, generate a plurality of extended vertices corresponding to each edge of interest, wherein the number of extended vertices of each edge of interest is proportional to the weight of the edge of interest;
[0011] Connecting the first vertex, the multiple extended vertices, and the second vertex of each edge of interest in sequence to obtain multiple extended edges connected in series in sequence, wherein the length of each extended edge in the multiple extended edges is equal;
[0012] The interested edge is replaced by the multiple extended edges connected in series in sequence of each interested edge to obtain a reconstructed road network graph.
[0013] Preferably, the process of generating a plurality of extended vertices corresponding to each edge of interest based on the weight of each edge of interest comprises:
[0014] The number of extended vertices n for each edge of interest is calculated using the following equation:
[0015] n=(w-1)×2,w∈N and w≥2
[0016] Where w is the weight of the edge of interest;
[0017] Based on the number of extended vertices of each edge of interest, a plurality of extended vertices corresponding to the edge of interest are generated.
[0018] Preferably, the process of replacing each extended edge in the extended test route with an edge of interest corresponding to the extended edge to obtain a test route corresponding to the road network graph includes:
[0019] The extended vertices corresponding to the same edge of interest in the extended test route are replaced two by two with the first vertex and the second vertex of the edge of interest according to the order of the extended vertices in the extended test route, so as to obtain a test route corresponding to the road network graph.
[0020] Preferably, the preset shortest path search algorithm is A* algorithm, Dijkstra algorithm, D* algorithm, LPA* algorithm or D*Lite algorithm.
[0021] A second aspect of the present application provides an autonomous driving road test method, comprising:
[0022] Get the road network map, starting point, end point and passing points;
[0023] Based on the road network diagram, the starting point, the end point and the waypoints, determine the test route using the test path planning method;
[0024] The vehicle is controlled to perform a road test using the test route to obtain a test result.
[0025] Preferably, it also includes:
[0026] When a new waypoint needs to be added, obtain the current point and the waypoints that have not been passed;
[0027] Taking the current point as a new starting point, constructing a new set of waypoints using the new waypoints and the waypoints that have not been passed, and determining a new test route using the test path planning method based on the road network graph, the new starting point, the end point and the new set of waypoints;
[0028] The vehicle is controlled to perform a road test using the new test route to obtain a test result.
[0029] Preferably, the test result includes the weight value of each edge in the road network graph, and the method further includes:
[0030] The weight of each edge in the road network graph is updated according to the weight value of each version in the road network graph.
[0031] Preferably, the process of updating the weight of each edge in the road network graph according to the weight value of each version in the road network graph includes:
[0032] Determine the weight value of each edge in the road network graph using a forgetting algorithm based on Newton's cooling formula according to the weight values of the most recent N versions in the road network graph;
[0033] Wherein, N is a preset natural number, and the Newton cooling formula is:
[0034] HH s =(H0-H s ) -kt
[0035] Among them, H s is the forgetting threshold, H0 is the initial value, k is the preset hyperparameter, and t is the time interval from a version to the current moment.
[0036] A third aspect of the present application provides a test path planning device, comprising:
[0037] An interesting edge splitting unit is used to split each interesting edge in the road network graph into a plurality of extended edges connected in series to obtain an extended road network graph, wherein the total length of the plurality of extended edges is equal to the length of the interesting edge;
[0038] An extended route determination unit is used to perform path planning based on the extended road network graph, in combination with a starting point, an end point and a set of waypoints, using a preset shortest path search algorithm to obtain an extended test route corresponding to the extended road network graph;
[0039] The test route determination unit is used to replace each extended edge in the extended test route with an edge of interest corresponding to the extended edge, so as to obtain a test route corresponding to the road network graph.
[0040] A fourth aspect of the present application provides an autonomous driving road test device, comprising:
[0041] A data acquisition unit, used to acquire a road network map, a starting point, an end point, and a passing point;
[0042] A path planning unit, configured to determine a test route based on the road network diagram, the starting point, the end point and the waypoints using the above-mentioned test path planning method;
[0043] The test execution unit is used to control the vehicle to perform a road test using the test route to obtain a test result.
[0044] A fifth aspect of the present application provides an autonomous driving road test device, including: a memory and a processor;
[0045] The memory is used to store programs;
[0046] The processor is used to execute the program to implement the above-mentioned test path planning method, or each step of the above-mentioned autonomous driving road testing method.
[0047] A sixth aspect of the present application provides a storage medium on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned test path planning method or each step of the above-mentioned autonomous driving road test method is implemented.
[0048] It can be known from the above technical solution that the present application first splits each edge of interest in the road network graph into multiple extended edges connected in series in sequence to obtain an extended road network graph. Among them, the total length of the multiple extended edges is equal to the length of the edge of interest, so that the length of each extended edge is less than the length of the edge of interest. It can be understood that the edges in the road network graph represent road sections, and the edges of interest represent the road sections of interest in the road test. Then, based on the extended road network graph, combined with the starting point, end point and waypoint set, a preset shortest path search algorithm is used for path planning to obtain an extended test route corresponding to the extended road network graph. It can be understood that since the edge of interest is split into multiple extended edges with a constant total length, when the total lengths of the searched candidate test routes are comparable, the shortest path search algorithm will preferably select the branch where the extended edge is located, so the obtained extended test route will definitely pass through each extended edge. Finally, each extended edge in the extended test route is replaced with an edge of interest corresponding to the extended edge, which is equivalent to the edge of interest being repeatedly used multiple times. The test route corresponding to the road network diagram obtained will also repeatedly pass through the edge of interest multiple times. In addition, due to the characteristics of the shortest path search algorithm, the total length of the searched test route is relatively short. Therefore, while completing the path planning, the present application achieves the effect of focusing on the interested section, and because the total length of the test route is relatively short, the test time can be used efficiently. BRIEF DESCRIPTION OF THE DRAWINGS
[0049] In order to more clearly illustrate the embodiments of the present application 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 embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.
[0050] Figure 1 A schematic diagram of a test path planning method disclosed in an embodiment of the present application;
[0051] Figure 2 An example of the road network diagram disclosed in the embodiment of the present application is provided;
[0052] Figure 3 An example of the extended road network diagram disclosed in the embodiment of the present application is provided;
[0053] Figure 4 Another extended road network diagram disclosed in the embodiment of the present application is illustrated;
[0054] Figure 5 A schematic diagram of an autonomous driving road test method disclosed in an embodiment of the present application;
[0055] Figure 6 A schematic diagram of a test path planning device disclosed in an embodiment of the present application;
[0056] Figure 7 A schematic diagram of an autonomous driving road test device disclosed in an embodiment of the present application;
[0057] Figure 8 This is a schematic diagram of an autonomous driving road test device disclosed in an embodiment of the present application. DETAILED DESCRIPTION
[0058] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0059] The following describes the test path planning method provided by the embodiment of the present application. Figure 1 The test path planning method provided in the embodiment of the present application may include the following steps:
[0060] Step S101, split each edge of interest in the road network graph into multiple extended edges connected in series to obtain an extended road network graph.
[0061] For each edge of interest, the total length of the multiple extended edges split from it is equal to the length of the edge of interest.
[0062] It can be understood that the road network graph is abstracted from the actual map. Usually, the road network graph is a directed graph. The vertices in the road network graph can be the intersections of the roads, and the edges in the road network graph represent the lane sections between the two intersections, and the direction of the edge represents the lane direction.
[0063] The interesting edge may be a lane section of interest in the autonomous driving test, for example, a lane section with frequent accidents, a lane section containing interesting scenes (such as involving complex intersections, involving sharp turns, involving lane merging), etc.
[0064] For example, Figure 2 In the road network graph shown in the figure, assuming that the edge {1, 2} formed by vertex 1 and vertex 2 is the edge of interest, then this edge can be split into extended edges {1, 6}, {6, 7} and {7, 2}. These three extended edges are connected in series in sequence, and the head and tail are connected to vertex 1 and vertex 2 respectively. The final extended road network graph is as follows: Figure 3 shown.
[0065] Step S102, based on the extended road network graph, combined with the starting point, the end point and the set of waypoints, a preset shortest path search algorithm is used to perform path planning to obtain an extended test route corresponding to the extended road network graph.
[0066] It can be understood that the waypoint set includes at least one waypoint, wherein the waypoint is a specified point that must be passed through, and the obtained extended test route must include these waypoints.
[0067] Starting from the starting point, along the edges of the graph to reach the end point, the path with the smallest sum of the lengths of the edges is called the shortest path. The shortest path search algorithm is an algorithm for solving the shortest path problem, which can usually include A* algorithm, Dijkstra algorithm, D* algorithm, LPA* algorithm or D*Lite algorithm.
[0068] Step S103, replacing each extended edge in the extended test route with an edge of interest corresponding to the extended edge, to obtain a test route corresponding to the road network graph.
[0069] For example, assuming that Figure 3In the extended road network diagram in step S102, the extended test route obtained is {1,4,1,6,7,2,3}, which is disassembled into edges connected in series. The extended test route can be represented as: {1,4}, {4,1}, {1,6}, {6,7}, {7,2}, {2,3}. Then, after replacing the extended edges {1,6}, {6,7}, {7,2} contained in the extended test route with the edge of interest {1,2} corresponding to the extended edge, the edges connected in series in the obtained test route are: {1,4}, {4,1}, {1,2}, {2,1}, {1,2}, {2,3}, and the test route corresponding to the road network diagram is {1,4,1,2,1,2,3}. It is easy to see that the test route passes through the edge of interest {1,2} twice, achieving the purpose of focusing on the road section of interest.
[0070] The present application first splits each edge of interest in the road network graph into multiple extended edges connected in series in sequence to obtain an extended road network graph. The total length of the multiple extended edges is equal to the length of the edge of interest, so that the length of each extended edge is less than the length of the edge of interest. It can be understood that the edges in the road network graph represent road sections, and the edges of interest represent road sections of interest in road testing. Then, based on the extended road network graph, combined with the starting point, end point and waypoint set, a preset shortest path search algorithm is used for path planning to obtain an extended test route corresponding to the extended road network graph. It can be understood that since the edge of interest is split into multiple extended edges with a constant total length, when the total lengths of the candidate test routes searched are comparable, the shortest path search algorithm will preferentially select the branch where the extended edge is located, so the obtained extended test route will definitely pass through each extended edge. Finally, each extended edge in the extended test route is replaced with an edge of interest corresponding to the extended edge, which is equivalent to the edge of interest being repeatedly used multiple times. The test route corresponding to the road network diagram obtained will also repeatedly pass through the edge of interest multiple times. In addition, due to the characteristics of the shortest path search algorithm, the total length of the searched test route is relatively short. Therefore, while completing the path planning, the present application achieves the effect of focusing on the interested section, and because the total length of the test route is relatively short, the test time can be used efficiently.
[0071] Furthermore, since the probability of accidents in road tests is generally low, there will not be a situation where a large number of roads have accidents at the same time, that is, the proportion of interesting edges in the road network graph is relatively small. Based on this situation, even if the splitting operation of the interesting edges (splitting each interesting edge in the road network graph into multiple extended edges) expands the number of edges in the road network graph, it has little effect on the complexity of finding the optimal path. Secondly, since the shortest path search algorithm itself considers the length of the edge in path planning, for the same road network graph, the path planning based only on the length after the split is consistent with the path planning based on the length and weight before the split. The path planning based on length and weight before the split has the following defects: First, each edge introduces a weight parameter, even if the edge has a weight of 0, it will greatly increase the spatial complexity of the path planning; secondly, after the weight parameter is introduced, it cannot adapt to the traditional path finding algorithm, and the path finding algorithm needs to be redesigned. The split-based mapping of the present application can obtain a simple directed graph without duplicate edges and self-loops. The particularity of this directed graph determines that it can adapt to most pathfinding algorithms, facilitating subsequent algorithm replacement and algorithm improvement.
[0072] In some embodiments of the present application, the interest level of the aforementioned interesting edge may be represented by the weight of the edge, and the higher the interest level, the greater the value of the weight.
[0073] Based on this, the process of splitting each edge of interest in the road network graph into a plurality of extended edges connected in series in the above step S101 to obtain a reconstructed road network graph may include:
[0074] S1, based on the weight of each edge of interest, generate a plurality of extended vertices corresponding to each edge of interest.
[0075] The number of extended vertices of each edge of interest is proportional to the weight of the edge of interest. Figure 2 The weight of the edge of interest {1, 2} in is 2, then the following can be generated based on this weight value: Figure 3 The two extended vertices shown are extended vertex 6 and extended vertex 7; assuming Figure 2 The weight of the edge of interest {1, 2} in is 3, then the following can be generated based on this weight value: Figure 4 The four extended vertices shown are: extended vertex 6, extended vertex 7, extended vertex 8 and extended vertex 9.
[0076] S2, sequentially connect the first vertex, multiple extended vertices, and the second vertex of each edge of interest to obtain multiple extended edges that are sequentially connected in series.
[0077] The first vertex and the second vertex are two endpoints of the edge of interest, and the length of each of the multiple extended edges is equal, so as to divide the length of the edge of interest into equal parts, thereby calculating the length of each extended edge.
[0078] S3, replacing each edge of interest with the multiple extended edges connected in series, to obtain a reconstructed road network graph.
[0079] In some embodiments of the present application, the process of generating a plurality of extended vertices corresponding to each edge of interest based on the weight of each edge of interest in S1 may include:
[0080] S11, the number n of extended vertices of each edge of interest is calculated using the following equation:
[0081] n=(w-1)×2,w∈N and w≥2
[0082] Where w is the weight of the edge of interest.
[0083] S12: Based on the number of extended vertices of each edge of interest, generate a plurality of extended vertices corresponding to the edge of interest.
[0084] It can be understood that, through the definition of S11, for each edge of interest, the number of expanded vertices obtained is an even number, which is convenient for subsequent further processing.
[0085] In some embodiments of the present application, the process of replacing each extended edge in the extended test route with an edge of interest corresponding to the extended edge in step S103 to obtain a test route corresponding to the road network graph may include:
[0086] The extended vertices corresponding to the same edge of interest in the extended test route are replaced two by two with the first vertex and the second vertex of the edge of interest according to the order of the extended vertices in the extended test route, so as to obtain a test route corresponding to the road network graph.
[0087] For example, first, the extended vertices corresponding to the same edge of interest in the extended test route can be formed into a queue according to the order of the extended vertices in the extended test route; then, with a step size of 2, this queue is cut into multiple sub-queues, each sub-queue includes 2 extended vertices; then, the extended vertices in each sub-queue are replaced with the first vertex and the second vertex of the edge of interest in turn, to obtain each replaced sub-queue, wherein, in the extended test route, the second vertex is connected to the head of the queue, and the first vertex is connected to the tail of the queue; finally, the replaced sub-queues are spliced into a new queue, and the new queue is used to replace the corresponding queue in the extended test route to obtain a test route corresponding to the road network graph.
[0088] For example, for Figure 3 The extended test route {1,4,1,6,7,2,3} shown in the figure involves only two extended vertices, namely extended vertex 6 and extended vertex 7. Then, extended vertex 6 and extended vertex 7 form a queue {6,7}. Since the queue has only two extended vertices, no further pruning is required. Since vertex 1 is connected to the head of the queue and vertex 2 is connected to the tail of the queue, vertex 1 is determined as the second vertex and vertex 2 is determined as the first vertex. The extended vertices in the queue {6,7} are replaced with the first vertex and the second vertex of the edge of interest in turn to obtain a new queue {2,1}. The new queue {2,1} is used to replace the corresponding queue {6,7} in the extended test route to obtain the test route {1,4,1,2,1,2,3} corresponding to the road network graph.
[0089] For Figure 4 The extended test route {1,4,1,6,7,8,9,2,3} shown in the figure involves 4 extended vertices, namely extended vertex 6, extended vertex 7, extended vertex 8 and extended vertex 9. Then, extended vertex 6, extended vertex 7, extended vertex 8 and extended vertex 9 form a queue {6,7,8,9}. With a step size of 2, the queue {6,7,8,9} is cut into 2 sub-queues: sub-queue {6,7} and sub-queue {8,9}. Since vertex 1 is connected to the head of the queue and vertex 2 is connected to the tail of the queue, vertex 1 is determined as the second vertex. Vertex 2 is determined as the first vertex; the extended vertices in the subqueues {6,7} and the subqueues {8,9} are replaced with the first vertex and the second vertex of the edge of interest respectively to obtain the replaced subqueues: subqueues {2,1} and subqueues {2,1}; the subqueues {2,1} and the subqueues {2,1} are concatenated into a new queue {2,1,2,1}, and the new queue {2,1,2,1} is used to replace the corresponding queues {6,7,8,9} in the extended test route to obtain the test route {1,4,1,2,1,2,1,2,3} corresponding to the road network graph.
[0090] Based on the test path planning methods provided in the above embodiments, this application also provides an autonomous driving road test method. Figure 5 The automatic driving road test method provided in the embodiment of the present application may include the following steps:
[0091] Step S201, obtaining a road network map, a starting point, an end point, and passing points.
[0092] For example, before each road test begins, an engineer uses a tool to mark the starting point, end point, and waypoints. After the engineer completes the marking, the path planning system obtains these data and the road network map and submits them to step S202 for path planning.
[0093] Step S202, determining a test route based on the road network diagram, the starting point, the end point and the passing points.
[0094] Specifically, the test route is determined using the test path planning method provided in any of the above embodiments.
[0095] Step S203, controlling the vehicle to perform a road test using the test route to obtain a test result.
[0096] During road testing, if engineers need to temporarily dispatch a vehicle to a new point of interest, they can add the new point of interest to the path planning system, which will then replan the path and send the result to the vehicle.
[0097] Based on this, in some embodiments of the present application, the autonomous driving road test method provided in the embodiments of the present application may further include:
[0098] Step S204, when a new waypoint needs to be added, the current point and the waypoints that have not been passed are obtained.
[0099] The current point is the next vertex that the vehicle is going to travel on the current driving path, which can also be understood as the vertex on the road network diagram that is closest to the current position of the vehicle along the current driving path. The waypoints that have not been passed yet mean that these waypoints need to be replanned into the new test route.
[0100] Step S205, re-planning the path based on the current point, new waypoints and waypoints that have not been passed, and determining a new test route.
[0101] Specifically, taking the current point as the new starting point, a new waypoint set is constructed using new waypoints and waypoints that have not yet been passed, and based on the original road network diagram, the new starting point, the end point and the new waypoint set, a new test route is determined using the test path planning method provided in any of the above embodiments.
[0102] Step S206, controlling the vehicle to perform a road test using the new test route to obtain a test result.
[0103] After the road test is completed, relevant data can be uploaded for use in the next road test. Among them, this data may include the accident situation of the road section. The probability of accident occurrence in the road section is the focus of the road test. The road section with a high probability of accident occurrence is also the road section of interest in the road test.
[0104] Among them, the degree of interest is related to the autonomous driving algorithm. For example, a certain version of the autonomous driving algorithm focuses on solving the problem of meeting vehicles, and the incidence of meeting accidents on certain road sections is relatively high, so the algorithm is more interested in these road sections with difficult meeting; a certain version of the autonomous driving algorithm focuses on solving the problem of merging lanes, and the incidence of merging lanes accidents on certain road sections is relatively high, so the algorithm is more interested in these road sections with merging lanes. Due to the iteration of the autonomous driving algorithm, the incidence of various accidents on the road section will change dynamically, so the weight of each road section needs to be dynamically maintained. For example, for each accident that occurs on the road section, it can be graded according to the severity of the accident, and the highest accident level is added to the weight of the road section.
[0105] Based on this, in some embodiments of the present application, the test result mentioned in the above step S203 may include the weight value of each edge in the road network graph, and the autonomous driving road test method may further include:
[0106] The weight of each edge in the road network graph is updated according to the weight values of each version in the road network graph.
[0107] In some embodiments of the present application, the process of updating the weight of each edge in the road network graph according to the weight values of each version in the road network graph may include:
[0108] According to the weight values of the most recent N versions of the road network graph, the weight value of each edge in the road network graph is determined using a forgetting algorithm based on Newton's cooling formula.
[0109] Wherein, N is a preset natural number.
[0110] Specifically, the Newton cooling formula can be expressed as:
[0111] HH s =(H0-H s ) -kt
[0112] Among them, H s is the forgetting threshold. When the H value in a certain version is less than this value, it is forgotten. H0 is the initial value, k is the preset hyperparameter, and t is the time interval from a certain version to the current moment.
[0113] For example, assume that for edge a, edge b, and edge c, the starting version is V1, and the subsequent versions are V2, V3, and V4, respectively. The time interval between each version is 1 time unit, the hyperparameter k is ln(2), and the forgetting threshold H is s is 1, and N is set to 3.
[0114] For edge a, the weights of each version are 2, 0, 0, 0 respectively; for edge b, the weights of each version are 4, 0, 0, 2 respectively; for edge c, the weights of each version are 0, 3, 3, 4 respectively.
[0115] When t=0, for the starting version V1, since there is no historical version, the weights of the three edges are 2, 4, and 0 respectively.
[0116] When t=1, for version V2, according to the above formula, for edge a:
[0117] H=1+(2-1) / 2=1.5>1
[0118] Since the influence value H of V1 is greater than the forgetting threshold, it is not forgotten. The weight value at this time is 1.5+0=1.5, which is rounded to 2.
[0119] For edge b:
[0120] H=1+(4-1) / 2=2.5>1
[0121] Since the influence value H of V1 is greater than the forgetting threshold, it is not forgotten. The weight value at this time is 2.5+0=2.5, which is rounded to 3.
[0122] For edge c:
[0123] H=1+(0-1) / 2=0.5<1
[0124] Since the influence value H of V1 is less than the forgetting threshold, it is forgotten and the weight is directly taken as 3.
[0125] By analogy, the weights of the three edges at each moment can be calculated.
[0126] The test path planning device provided in an embodiment of the present application is described below. The test path planning device described below and the test path planning method described above can be referenced to each other.
[0127] See also Figure 6 The test path planning device provided in the embodiment of the present application may include:
[0128] An interested edge splitting unit 21 is used to split each interested edge in the road network graph into a plurality of extended edges connected in series to obtain an extended road network graph, wherein the total length of the plurality of extended edges is equal to the length of the interested edge;
[0129] The extended route determination unit 22 is used to perform path planning based on the extended road network graph, in combination with the starting point, the end point and the set of waypoints, using a preset shortest path search algorithm to obtain an extended test route corresponding to the extended road network graph;
[0130] The test route determining unit 23 is used to replace each extended edge in the extended test route with an edge of interest corresponding to the extended edge, so as to obtain a test route corresponding to the road network graph.
[0131] In some embodiments of the present application, the interest level of the interesting edge is determined by the weight of the edge. The interesting edge splitting unit 21 splits each interesting edge in the road network graph into a plurality of extended edges connected in series in sequence to obtain a reconstructed road network graph, which may include:
[0132] Based on the weight of each edge of interest, generate a plurality of extended vertices corresponding to each edge of interest, wherein the number of extended vertices of each edge of interest is proportional to the weight of the edge of interest;
[0133] Connecting the first vertex, the multiple extended vertices, and the second vertex of each edge of interest in sequence to obtain multiple extended edges connected in series in sequence, wherein the length of each extended edge in the multiple extended edges is equal;
[0134] The interested edge is replaced by the multiple extended edges connected in series in sequence of each interested edge to obtain a reconstructed road network graph.
[0135] In some embodiments of the present application, the process of generating a plurality of extended vertices corresponding to each edge of interest based on the weight of each edge of interest by the interest edge splitting unit 21 may include:
[0136] The number of extended vertices n for each edge of interest is calculated using the following equation:
[0137] n=(w-1)×2,w∈N and w≥2
[0138] Where w is the weight of the edge of interest;
[0139] Based on the number of extended vertices of each edge of interest, a plurality of extended vertices corresponding to the edge of interest are generated.
[0140] In some embodiments of the present application, the process in which the test route determining unit 23 replaces each extended edge in the extended test route with an edge of interest corresponding to the extended edge to obtain a test route corresponding to the road network graph may include:
[0141] The extended vertices corresponding to the same edge of interest in the extended test route are replaced two by two with the first vertex and the second vertex of the edge of interest according to the order of the extended vertices in the extended test route, so as to obtain a test route corresponding to the road network graph.
[0142] In some embodiments of the present application, the preset shortest path search algorithm is A* algorithm, Dijkstra algorithm, D* algorithm, LPA* algorithm or D*Lite algorithm
[0143] The test path planning device provided in the embodiment of the present application can be applied to a test path planning device, such as a computer. Optionally, Figure 7 The hardware structure diagram of the test path planning device is shown in FIG. Figure 7 The hardware structure of the test path planning device may include: at least one processor 31 , at least one communication interface 32 , at least one memory 33 and at least one communication bus 34 .
[0144] In the embodiment of the present application, the number of the processor 31, the communication interface 32, the memory 33, and the communication bus 34 is at least one, and the processor 31, the communication interface 32, and the memory 33 communicate with each other through the communication bus 34;
[0145] The processor 31 may be a central processing unit CPU, or an application-specific integrated circuit ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present application, etc.;
[0146] The memory 33 may include a high-speed RAM memory, and may also include a non-volatile memory, such as at least one disk memory;
[0147] The memory 33 stores a program, and the processor 31 can call the program stored in the memory 33, and the program is used to:
[0148] Splitting each edge of interest in the road network graph into a plurality of extended edges connected in series, to obtain an extended road network graph, wherein the total length of the plurality of extended edges is equal to the length of the edge of interest;
[0149] Based on the extended road network graph, in combination with the starting point, the end point and the set of waypoints, a preset shortest path search algorithm is used to perform path planning to obtain an extended test route corresponding to the extended road network graph;
[0150] Each extended edge in the extended test route is replaced with an edge of interest corresponding to the extended edge, so as to obtain a test route corresponding to the road network graph.
[0151] Optionally, the detailed functions and extended functions of the program may refer to the above description.
[0152] The embodiment of the present application further provides a storage medium, which may store a program suitable for execution by a processor, wherein the program is used to:
[0153] Splitting each edge of interest in the road network graph into a plurality of extended edges connected in series, to obtain an extended road network graph, wherein the total length of the plurality of extended edges is equal to the length of the edge of interest;
[0154] Based on the extended road network graph, in combination with the starting point, the end point and the set of waypoints, a preset shortest path search algorithm is used to perform path planning to obtain an extended test route corresponding to the extended road network graph;
[0155] Each extended edge in the extended test route is replaced with an edge of interest corresponding to the extended edge, so as to obtain a test route corresponding to the road network graph.
[0156] Optionally, the detailed functions and extended functions of the program may refer to the above description.
[0157] The following is a description of the autonomous driving road testing device provided in an embodiment of the present application. The autonomous driving road testing device described below and the autonomous driving road testing method described above can be referenced to each other.
[0158] See also Figure 8 The automatic driving road test device provided in the embodiment of the present application may include:
[0159] The data acquisition unit 10 is used to acquire the road network map, the starting point, the end point and the passing points.
[0160] The path planning unit 20 is used to determine the test route by using a test path planning method based on the road network diagram, the starting point, the end point and the passing points.
[0161] Among them, the test path planning method is the test path planning method provided by any of the above embodiments.
[0162] The test execution unit 30 is used to control the vehicle to perform a road test using the test route to obtain a test result.
[0163] In some embodiments of the present application, the autonomous driving road test device may further include a waypoint adding unit, wherein the waypoint adding unit is used to:
[0164] When a new waypoint needs to be added, obtain the current point and the waypoints that have not been passed;
[0165] Taking the current point as a new starting point, constructing a new waypoint set using the new waypoint and the waypoints that have not been passed, and determining a new test route using the test path planning method provided in any of the above embodiments based on the road network diagram, the new starting point, the end point and the new waypoint set;
[0166] The vehicle is controlled to perform a road test using the new test route to obtain a test result.
[0167] In some embodiments of the present application, the test result includes a weight value of each edge in the road network graph, and the autonomous driving road test device may further include a weight updating unit, wherein the weight updating unit is used to:
[0168] The weight of each edge in the road network graph is updated according to the weight value of each version in the road network graph.
[0169] In some embodiments of the present application, the process of updating the weight of each edge in the road network graph according to the weight value of each version in the road network graph by the weight updating unit may include:
[0170] Determine the weight value of each edge in the road network graph using a forgetting algorithm based on Newton's cooling formula according to the weight values of the most recent N versions in the road network graph;
[0171] Wherein, N is a preset natural number, and the Newton cooling formula is:
[0172] HH s =(H0-H s ) -kt
[0173] Among them, H s is the forgetting threshold, H0 is the initial value, k is the preset hyperparameter, and t is the time interval from a version to the current moment.
[0174] The autonomous driving road test device provided in the embodiment of the present application can be applied to autonomous driving road test equipment, such as a computer, etc. Optionally, Figure 7 The hardware structure diagram of the autonomous driving road test equipment is shown in FIG. Figure 7 The hardware structure of the autonomous driving road test device may include: at least one processor 31, at least one communication interface 32, at least one memory 33 and at least one communication bus 34.
[0175] In the embodiment of the present application, the number of the processor 31, the communication interface 32, the memory 33, and the communication bus 34 is at least one, and the processor 31, the communication interface 32, and the memory 33 communicate with each other through the communication bus 34;
[0176] The processor 31 may be a central processing unit CPU, or an application-specific integrated circuit ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present application, etc.;
[0177] The memory 33 may include a high-speed RAM memory, and may also include a non-volatile memory, such as at least one disk memory;
[0178] The memory 33 stores a program, and the processor 31 can call the program stored in the memory 33, and the program is used to:
[0179] Get the road network map, starting point, end point and passing points;
[0180] Based on the road network diagram, the starting point, the end point and the waypoints, determine a test route using a test path planning method provided in any of the above embodiments;
[0181] The vehicle is controlled to perform a road test using the test route to obtain a test result.
[0182] Optionally, the detailed functions and extended functions of the program may refer to the above description.
[0183] The embodiment of the present application further provides a storage medium, which may store a program suitable for execution by a processor, wherein the program is used to:
[0184] Get the road network map, starting point, end point and passing points;
[0185] Based on the road network diagram, the starting point, the end point and the waypoints, determine a test route using a test path planning method provided in any of the above embodiments;
[0186] The vehicle is controlled to perform a road test using the test route to obtain a test result.
[0187] Optionally, the detailed functions and extended functions of the program may refer to the above description.
[0188] In summary:
[0189] The present application first splits each edge of interest in the road network graph into multiple extended edges connected in series in sequence to obtain an extended road network graph. The total length of the multiple extended edges is equal to the length of the edge of interest, so that the length of each extended edge is less than the length of the edge of interest. It can be understood that the edges in the road network graph represent road sections, and the edges of interest represent road sections of interest in road testing. Then, based on the extended road network graph, combined with the starting point, end point and waypoint set, a preset shortest path search algorithm is used for path planning to obtain an extended test route corresponding to the extended road network graph. It can be understood that since the edge of interest is split into multiple extended edges with a constant total length, when the total lengths of the candidate test routes searched are comparable, the shortest path search algorithm will preferentially select the branch where the extended edge is located, so the obtained extended test route will definitely pass through each extended edge. Finally, each extended edge in the extended test route is replaced with an edge of interest corresponding to the extended edge, which is equivalent to the edge of interest being repeatedly used multiple times. The test route corresponding to the road network diagram obtained will also repeatedly pass through the edge of interest multiple times. In addition, due to the characteristics of the shortest path search algorithm, the total length of the searched test route is relatively short. Therefore, while completing the path planning, the present application achieves the effect of focusing on the interested section, and because the total length of the test route is relatively short, the test time can be used efficiently.
[0190] Furthermore, since the probability of accidents in road tests is generally low, there will not be a situation where a large number of roads have accidents at the same time, that is, the proportion of interesting edges in the road network graph is relatively small. Based on this situation, even if the splitting operation of the interesting edges (splitting each interesting edge in the road network graph into multiple extended edges) expands the number of edges in the road network graph, it has little effect on the complexity of finding the optimal path. Secondly, since the shortest path search algorithm itself considers the length of the edge in path planning, for the same road network graph, the path planning based only on the length after the split is consistent with the path planning based on the length and weight before the split. The path planning based on length and weight before the split has the following defects: First, each edge introduces a weight parameter, even if the edge has a weight of 0, it will greatly increase the spatial complexity of the path planning; secondly, after the weight parameter is introduced, it cannot adapt to the traditional path finding algorithm, and the path finding algorithm needs to be redesigned. The split-based mapping of the present application can obtain a simple directed graph without duplicate edges and self-loops. The particularity of this directed graph determines that it can adapt to most pathfinding algorithms, facilitating subsequent algorithm replacement and algorithm improvement.
[0191] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.
[0192] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The various embodiments can be combined as needed, and the same or similar parts can refer to each other.
[0193] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A test path planning method, characterized in that: include: Splitting each edge of interest in the road network graph into a plurality of extended edges connected in series, to obtain an extended road network graph, wherein the total length of the plurality of extended edges is equal to the length of the edge of interest; Based on the extended road network graph, in combination with the starting point, the end point and the set of waypoints, a preset shortest path search algorithm is used to perform path planning to obtain an extended test route corresponding to the extended road network graph; Replacing each extended edge in the extended test route with an edge of interest corresponding to the plurality of extended edges, to obtain a test route corresponding to the road network graph; The interest level of the edge of interest is determined by the weight of the edge. The process of splitting each edge of interest in the road network graph into a plurality of extended edges connected in series to obtain a reconstructed road network graph includes: Based on the weight of each edge of interest, generate a plurality of extended vertices corresponding to each edge of interest, wherein the number of extended vertices of each edge of interest is proportional to the weight of the edge of interest; Connecting the first vertex, the multiple extended vertices, and the second vertex of each edge of interest in sequence to obtain multiple extended edges connected in series in sequence, wherein the length of each extended edge in the multiple extended edges is equal; Replacing each edge of interest with the multiple extended edges connected in series in sequence to obtain a reconstructed road network graph; The process of generating a plurality of extended vertices corresponding to each edge of interest based on the weight of each edge of interest includes: The number of extended vertices n for each edge of interest is calculated using the following equation: n=(w-1)×2,w∈N and w≥2 Where w is the weight of the edge of interest; Based on the number of extended vertices of each edge of interest, a plurality of extended vertices corresponding to the edge of interest are generated.
2. The method according to claim 1, characterized in that The process of replacing each extended edge in the extended test route with an edge of interest corresponding to the plurality of extended edges to obtain a test route corresponding to the road network graph includes: The extended vertices corresponding to the same edge of interest in the extended test route are replaced two by two with the first vertex and the second vertex of the edge of interest according to the order of the extended vertices in the extended test route, so as to obtain a test route corresponding to the road network graph.
3. An automatic driving road test method, characterized in that: include: Get the road network map, starting point, end point and passing points; Based on the road network diagram, the starting point, the end point and the waypoints, a test route is determined using the test path planning method according to any one of claims 1 to 2; The vehicle is controlled to perform a road test using the test route to obtain a test result.
4. The method according to claim 3, characterized in that Also includes: When a new waypoint needs to be added, obtain the current point and the waypoints that have not been passed; Taking the current point as a new starting point, constructing a new set of waypoints using the new waypoints and the waypoints that have not been passed, and determining a new test route using the test path planning method according to any one of claims 1 to 2 based on the road network diagram, the new starting point, the end point and the new set of waypoints; The vehicle is controlled to perform a road test using the new test route to obtain a test result.
5. The method according to claim 4, characterized in that The test result includes the weight value of each edge in the road network graph, and the method further includes: The weight of each edge in the road network graph is updated according to the weight value of each version in the road network graph.
6. The method according to claim 5, characterized in that The process of updating the weight of each edge in the road network graph according to the weight value of each version in the road network graph includes: Determine the weight value of each edge in the road network graph using a forgetting algorithm based on Newton's cooling formula according to the weight values of the most recent N versions in the road network graph; Wherein, N is a preset natural number, and the Newton cooling formula is: H.H. s =(H0-H s )e -kt Among them, H s is the forgetting threshold, H0 is the initial value, k is the preset hyperparameter, and t is the time interval from a version to the current moment.
7. A test path planning device, characterized in that: include: An interesting edge splitting unit is used to split each interesting edge in the road network graph into a plurality of extended edges connected in series to obtain an extended road network graph, wherein the total length of the plurality of extended edges is equal to the length of the interesting edge; An extended route determination unit is used to perform path planning based on the extended road network graph, in combination with a starting point, an end point and a set of waypoints, using a preset shortest path search algorithm to obtain an extended test route corresponding to the extended road network graph; A test route determination unit, configured to replace each extended edge in the extended test route with an edge of interest corresponding to the plurality of extended edges, to obtain a test route corresponding to the road network graph; The interest level of the edge of interest is determined by the weight of the edge. The process of splitting each edge of interest in the road network graph into a plurality of extended edges connected in series to obtain a reconstructed road network graph includes: Based on the weight of each edge of interest, generate a plurality of extended vertices corresponding to each edge of interest, wherein the number of extended vertices of each edge of interest is proportional to the weight of the edge of interest; Connecting the first vertex, the multiple extended vertices, and the second vertex of each edge of interest in sequence to obtain multiple extended edges connected in series in sequence, wherein the length of each extended edge in the multiple extended edges is equal; Replacing each edge of interest with the multiple extended edges connected in series in sequence to obtain a reconstructed road network graph; The process of generating a plurality of extended vertices corresponding to each edge of interest based on the weight of each edge of interest includes: The number of extended vertices n for each edge of interest is calculated using the following equation: n=(w-1)×2,w∈N and w≥2 Where w is the weight of the edge of interest; Based on the number of extended vertices of each edge of interest, a plurality of extended vertices corresponding to the edge of interest are generated.
8. An automatic driving road test device, characterized in that: include: A data acquisition unit, used to acquire a road network map, a starting point, an end point, and a passing point; A path planning unit, configured to determine a test route based on the road network diagram, the starting point, the end point and the waypoints using the test path planning method according to any one of claims 1 to 2; The test execution unit is used to control the vehicle to perform a road test using the test route to obtain a test result.
9. An automatic driving road test device, characterized in that: include: Memory and processor; The memory is used to store programs; The processor is used to execute the program to implement the test path planning method according to any one of claims 1 to 2, or the various steps of the automatic driving road testing method according to any one of claims 3 to 6.
10. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, each step of the test path planning method according to any one of claims 1 to 2 or the autonomous driving road testing method according to any one of claims 3 to 6 is implemented.
