Trajectory determination method, apparatus, device, and vehicle

Abstract

Embodiments of the present application provide a trajectory determination method, device, equipment and vehicle. The method comprises: obtaining at least two initial trajectories of a vehicle; determining rule information corresponding to each trajectory decision dimension in N trajectory decision dimensions; determining reference rule information from the rule information according to the priority of the rule information, and comparing each initial trajectory according to the reference rule information to obtain at least one candidate trajectory; and determining an optimal trajectory of the vehicle from the initial trajectories according to the at least one candidate trajectory. The method can realize the transparency of trajectory determination logic and reduce the complexity of parameter tuning.

Description

Technical Field

[0001] This application relates to the field of autonomous driving technology, and in particular to a trajectory determination method, apparatus, device, and vehicle. Background Technology

[0002] In complex urban roads, highways, parking lots, and other scenarios, vehicles need to plan multiple initial trajectories in real time and select the optimal trajectory to ensure safety and traffic efficiency.

[0003] In existing technologies, the commonly used trajectory decision-making scheme is based on the weighted cost function method. This method designs an independent cost function for each decision dimension and assigns a fixed weight to each cost function. Then, the cost value of each trajectory in all dimensions is weighted and summed to obtain the total cost. Finally, the trajectory with the minimum total cost is selected as the optimal solution.

[0004] However, in the above methods, manual adjustment of weights requires repeated balancing of multiple dimensions, making weight adjustment complex and difficult to maintain. Summary of the Invention

[0005] This application provides a trajectory determination method, apparatus, device, and vehicle, which can make the trajectory determination logic transparent and reduce the complexity of parameter tuning.

[0006] In a first aspect, embodiments of this application provide a trajectory determination method, including:

[0007] Obtain at least two initial trajectories for the vehicle;

[0008] Determine the rule information corresponding to each of the N trajectory decision dimensions; wherein the rule information has a priority, and N is a positive integer greater than or equal to 2;

[0009] Reference rule information is determined from each rule information according to the priority of the rule information, and each initial trajectory is compared pairwise according to the reference rule information to obtain at least one candidate trajectory.

[0010] Based on the at least one candidate trajectory, the optimal trajectory of the vehicle is determined among the initial trajectories.

[0011] In one possible implementation, the step of determining reference rule information from each set of rule information according to the priority of the rule information, and comparing each initial trajectory pairwise according to the reference rule information to obtain at least one candidate trajectory, includes:

[0012] The rule information is sorted in descending order of priority.

[0013] For any trajectory combination consisting of two initial trajectories, let i=1, and use the i-th rule information as reference rule information to compare the two initial trajectories in the trajectory combination, and determine the candidate trajectory in the trajectory combination based on the comparison result.

[0014] If the candidate trajectory in the trajectory combination cannot be determined based on the comparison result, then let i = i + 1, and return to the step of using the i-th rule information as reference trajectory rule information to compare the two initial trajectories in the trajectory combination, and determine the candidate trajectory in the trajectory combination based on the comparison result.

[0015] In one possible implementation, comparing two initial trajectories in the trajectory combination and determining candidate trajectories in the trajectory combination based on the comparison result includes:

[0016] Based on the reference rule information, calculate the trajectory cost value corresponding to the two initial trajectories in the trajectory combination respectively;

[0017] If the difference between the trajectory values ​​of two initial trajectories is greater than the comparison threshold, then the initial trajectory with the lower corresponding trajectory value in the trajectory combination is determined as the candidate trajectory of the trajectory combination.

[0018] In one possible implementation, the method further includes:

[0019] When i=N and the candidate trajectory of the trajectory combination cannot be determined based on the comparison results, the comparison threshold is adjusted.

[0020] Return to the step of setting i=1, using the i-th rule information as reference trajectory rule information to compare the two initial trajectories in the trajectory combination, and determining the candidate trajectory in the trajectory combination based on the comparison result.

[0021] In one possible implementation, determining the optimal trajectory of the vehicle based on the at least one candidate trajectory includes:

[0022] A directed graph is constructed based on the at least one candidate trajectory and each of the initial trajectories; wherein each vertex in the directed graph is used to represent each of the initial trajectories, and each arrow in the arrow pointing table between two vertices points to a candidate trajectory between the corresponding two initial trajectories.

[0023] The optimal trajectory of the vehicle is determined from each of the initial trajectories based on the in-degree of each vertex of the directed graph.

[0024] In one possible implementation, determining the optimal trajectory of the vehicle from each of the initial trajectories based on the in-degree of each vertex of the directed graph includes:

[0025] The vertex with the largest in-degree is determined, and the initial trajectory corresponding to the vertex with the largest in-degree is determined as the optimal trajectory of the vehicle.

[0026] In one possible implementation, if the number of vertices with the highest in-degree is greater than 1, the method further includes:

[0027] The initial trajectory corresponding to the vertex with the largest in-degree is taken as the new initial trajectory, and the process is returned to execute the steps of determining reference rule information from each rule information according to the priority of the rule information, and comparing each initial trajectory pairwise according to the reference rule information to obtain at least one candidate trajectory.

[0028] In one possible implementation, determining the rule information corresponding to each of the N trajectory decision dimensions includes:

[0029] Based on real-time scene information, determine the rule information corresponding to each trajectory decision dimension and / or the priority of each rule information.

[0030] Secondly, embodiments of this application provide a trajectory determination device, comprising:

[0031] The acquisition module is used to acquire at least two initial trajectories of the vehicle;

[0032] The first determining module is used to determine the rule information corresponding to each of the N trajectory decision dimensions; wherein the rule information has a priority, and N is a positive integer greater than or equal to 2;

[0033] The comparison module is used to determine reference rule information from each rule information according to the priority of the rule information, and to compare each initial trajectory pairwise according to the reference rule information to obtain at least one candidate trajectory;

[0034] The second determining module is used to determine the optimal trajectory of the vehicle among each initial trajectory based on the at least one candidate trajectory.

[0035] In one possible implementation, the comparison module is specifically used for: sorting the rule information in descending order of priority; for any trajectory combination consisting of two initial trajectories, setting i=1, using the i-th rule information as reference rule information to compare the two initial trajectories in the trajectory combination, and determining the candidate trajectory in the trajectory combination based on the comparison result; if the candidate trajectory in the trajectory combination cannot be determined based on the comparison result, setting i=i+1, and returning to the step of using the i-th rule information as reference trajectory rule information to compare the two initial trajectories in the trajectory combination, and determining the candidate trajectory in the trajectory combination based on the comparison result.

[0036] In one possible implementation, the comparison module is specifically used to: calculate the trajectory cost value corresponding to the two initial trajectories in the trajectory combination according to the reference rule information; if the difference between the trajectory cost values ​​of the two initial trajectories is greater than the comparison threshold, then the initial trajectory with the lower trajectory cost value in the trajectory combination is determined as the candidate trajectory of the trajectory combination.

[0037] In one possible implementation, the comparison module is further specifically used for: adjusting the comparison threshold when i=N and the candidate trajectory of the trajectory combination cannot be determined based on the comparison result; returning to the step of setting i=1, using the i-th rule information as reference trajectory rule information to compare the two initial trajectories in the trajectory combination, and determining the candidate trajectory in the trajectory combination based on the comparison result.

[0038] In one possible implementation, the second determining module is specifically configured to: construct a directed graph based on the at least one candidate trajectory and each of the initial trajectories; wherein each vertex in the directed graph is used to represent each of the initial trajectories, and each arrow in the arrow pointing table between two vertices points to a candidate trajectory between the corresponding two initial trajectories; and determine the optimal trajectory of the vehicle from each of the initial trajectories based on the in-degree of each vertex in the directed graph.

[0039] In one possible implementation, the second determining module is specifically used to: determine the vertex with the largest in-degree, and determine the initial trajectory corresponding to the vertex with the largest in-degree as the optimal trajectory of the vehicle.

[0040] In one possible implementation, when the number of vertices with the largest in-degree is greater than 1, the second determining module is further specifically used to: take the initial trajectory corresponding to the vertex with the largest in-degree as the new initial trajectory, and return to execute the step of determining reference rule information from each rule information according to the priority of the rule information, and comparing each initial trajectory pairwise according to the reference rule information to obtain at least one candidate trajectory.

[0041] In one possible implementation, the first determining module is specifically used to: determine the rule information corresponding to each trajectory decision dimension and / or the priority of each rule information based on real-time scene information.

[0042] Thirdly, embodiments of this application provide an electronic device, including: a memory and a processor;

[0043] The memory stores computer-executed instructions;

[0044] The processor executes computer execution instructions stored in the memory, causing the processor to perform the first aspect and / or various possible implementations of the first aspect as described above.

[0045] Fourthly, embodiments of this application provide a vehicle that includes the electronic equipment described in the third aspect.

[0046] Fifthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the first aspect and / or various possible implementations of the first aspect.

[0047] In a sixth aspect, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the first aspect and / or various possible implementations of the first aspect.

[0048] The trajectory determination method, apparatus, device, and vehicle provided in this application determine the rule information corresponding to each trajectory decision dimension and determine the reference rule information according to the priority of each rule information. This allows for pairwise comparison of multiple initial trajectories to obtain at least one candidate trajectory, thereby determining the optimal trajectory of the vehicle from each initial trajectory. Furthermore, by adopting a trajectory decision logic that prioritizes rules, the transparency of the trajectory determination logic can be achieved and the complexity of parameter tuning can be reduced. Attached Figure Description

[0049] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0050] Figure 1 This application provides an illustration of an application scenario.

[0051] Figure 2 A schematic flowchart illustrating a trajectory determination method provided in an embodiment of this application;

[0052] Figure 3A flowchart illustrating another trajectory determination method provided in an embodiment of this application;

[0053] Figure 4 An example diagram of a rulebook provided for an embodiment of this application;

[0054] Figure 5 A flowchart illustrating an autonomous driving trajectory decision-making method based on a rulebook, provided as an embodiment of this application;

[0055] Figure 6 A schematic diagram of a pairwise comparison process for a rulebook is provided as an embodiment of this application;

[0056] Figure 7 Another schematic diagram of the pairwise comparison process of rulebooks provided for embodiments of this application;

[0057] Figure 8 A rulebook directed graph provided for embodiments of this application;

[0058] Figure 9 A schematic diagram of a rulebook forming a loop is provided as an embodiment of this application;

[0059] Figure 10 This is a schematic diagram of the structure of a trajectory determination device provided in an embodiment of this application;

[0060] Figure 11 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0061] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0062] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0063] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use, processing, transmission, provision, disclosure, and application of the relevant data all comply with relevant laws, regulations, and standards, necessary confidentiality measures have been taken, they do not violate public order and good morals, and corresponding operation portals are provided for users to choose to authorize or refuse.

[0064] Furthermore, the technical solution involved in this application, which involves big data analysis of user information (including but not limited to personal biometrics, identity data, consumption data, asset data, electronic terminal operation data, etc.) and the use of artificial intelligence technology for automated decision-making, and makes decisions that have a significant impact on personal rights based on the results of automated decision-making, provides users with corresponding operation entry points for users to choose to agree to or reject the results of automated decision-making; if the user chooses to reject, the process will proceed to the expert decision-making process.

[0065] Figure 1 This application provides an illustration of an application scenario, such as... Figure 1 As shown, the specific application scenario of this application is as follows: In current autonomous driving systems, the trajectory planning module typically generates multiple alternative trajectories (i.e., "multimodal trajectories") for the vehicle to drive. Subsequently, the trajectory decision module needs to select the most reasonable and safest trajectory from these alternative trajectories and output it to the downstream control module for execution.

[0066] As can be seen from the above scenarios, the method based on the weighted cost function has technical problems such as complex weight adjustment and difficulty in maintenance.

[0067] The trajectory determination method provided in this application uses a multi-dimensional logic of trajectory decision-making based on rule priority ranking, and introduces dynamic adjustment rules and directed graphs for trajectory decision-making, which can make the trajectory determination logic transparent and reduce the complexity of parameter tuning.

[0068] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0069] Figure 2 This is a flowchart illustrating a trajectory determination method provided in an embodiment of this application, as shown below. Figure 2 As shown, the method includes:

[0070] 201. Obtain at least two initial trajectories for the vehicle.

[0071] For example, the execution subject of this embodiment can be an electronic device, hereinafter referred to as the device. The device can be deployed on the vehicle or in the cloud. Based on the device, the path planning algorithm of the trajectory planning module can generate multiple initial trajectories for the vehicle to travel based on the current location of the vehicle's origin and destination.

[0072] In one example, based on the trajectory planning module, multiple initial trajectories are generated according to the vehicle's origin and destination locations, as well as the traffic control situation within the road area where the vehicle is currently located. Each initial trajectory may contain information such as travel time and travel distance.

[0073] 202. Determine the rule information corresponding to each of the N trajectory decision dimensions; where the rule information has priority, and N is a positive integer greater than or equal to 2.

[0074] For example, the device invokes N preset trajectory decision dimensions to consider, where N is a positive integer greater than or equal to 2. For each trajectory decision dimension, an independent rule is configured, resulting in the corresponding rule information. Each rule is a judgment condition used to compare the merits of two initial trajectories in a certain dimension. Each rule has a priority, which represents the degree of priority for using the rule information to compare the initial trajectories; a higher priority indicates a higher degree of priority in use.

[0075] In one implementation, N trajectory decision dimensions include safety, comfort, traffic rule compliance, and efficiency. For the safety dimension, corresponding safety rules are configured, including determining whether trajectory A is safer than trajectory B. For the comfort dimension, corresponding comfort rules are configured, including determining whether trajectory A is more comfortable than trajectory B. For the traffic rule compliance dimension, corresponding traffic rule compliance rules are configured, including determining whether trajectory A conforms better to traffic signs, lane markings, and other rules than trajectory B. For the efficiency dimension, corresponding efficiency rules are configured, including determining whether the driving efficiency of trajectory A is higher than that of trajectory B.

[0076] 203. Determine reference rule information from each rule information according to the priority of the rule information, and compare each initial trajectory pairwise according to the reference rule information to obtain at least one candidate trajectory.

[0077] For example, reference rule information is determined from the rule information based on the priority of each rule information. Based on the judgment conditions corresponding to the reference rule information, all initial trajectories are compared pairwise to obtain the results of the pairwise comparisons. Based on all the results of the pairwise comparisons, at least one candidate trajectory is determined from each initial trajectory.

[0078] In one example, based on the priority of each rule, the rule with the highest priority is determined as the reference rule for pairwise comparisons of all initial trajectories. For example, the rule information corresponding to N trajectory decision dimensions includes safety rules, comfort rules, and efficiency rules, with the safety rule having the highest priority. For the four initial trajectories A, B, C, and D, they are divided into six combinations: A and B, A and C, A and D, B and C, B and D, and C and D. According to the judgment conditions corresponding to the safety rule, A and B are compared, and simultaneously, A and C, A and D, B and C, B and D, and C and D are compared, yielding the comparison results for A and B, A and C, A and D, B and C, B and D, and C and D, respectively.

[0079] In one example, reference rule information is determined from the various rule information based on their priority and in conjunction with the current traffic flow status (e.g., whether the current traffic flow is smooth, slow, or stagnant).

[0080] 204. Based on at least one candidate trajectory, determine the optimal trajectory for the vehicle among the initial trajectories.

[0081] For example, each candidate trajectory is analyzed based on a preset trajectory decision strategy (e.g., a trajectory decision strategy based on driving time or a trajectory decision strategy based on driving distance) to obtain the optimal trajectory for the vehicle.

[0082] In one example, a trajectory decision-making strategy based on travel time determines the candidate trajectory with the shortest expected travel time from all candidate trajectories as the vehicle's optimal trajectory; alternatively, a trajectory decision-making strategy based on travel distance determines the candidate trajectory with the shortest expected travel distance from all candidate trajectories as the vehicle's optimal trajectory; or, a weighted calculation is performed on the expected travel time and expected travel distance of each candidate trajectory to obtain a comprehensive result, and the candidate trajectory with the best comprehensive result is determined as the vehicle's optimal trajectory.

[0083] This embodiment provides a trajectory determination method that changes the multi-dimensional logic of trajectory decision-making from "weighted summation" to "rule priority sorting". By replacing the traditional weight parameter adjustment with a hierarchical decision-making mechanism, the decision logic design can be completed simply by adjusting the rule order, avoiding the "seesaw effect" and realizing the transparency of the decision logic and the simplification of parameter adjustment.

[0084] Figure 3 A flowchart illustrating another trajectory determination method provided in this application embodiment is shown below. Figure 3 As shown, the method includes:

[0085] 301. Obtain at least two initial trajectories for the vehicle.

[0086] For example, this step can be referred to as step 201, which will not be repeated here.

[0087] 302. Based on real-time scene information, determine the rule information corresponding to each trajectory decision dimension and / or the priority of each rule information.

[0088] For example, the device acquires real-time scene information of the vehicle's environment through an onboard perception module (e.g., sensors), including the scene type (e.g., structured road or urban village road). Based on this real-time scene information, rule information corresponding to each trajectory decision dimension can be configured, including adding / deleting / adjusting rule information. And / or, based on the real-time scene information, the priority of the rule information corresponding to each trajectory decision dimension can be configured, including adjusting the priority of the configured rule information.

[0089] In one example, Figure 4 An example diagram of a rulebook provided for embodiments of this application, such as... Figure 4 As shown, this rulebook includes pre-configured rule information corresponding to N trajectory decision dimensions, including safety rules, comfort rules, traffic rule compliance rules, and efficiency rules. Based on the priority of each rule, the order of the rule information in the rulebook can be: safety rules -> traffic rule compliance rules -> comfort rules -> efficiency rules. This order reflects the decision-making logic of "safety first, efficiency last." By placing safety as the highest priority, the mechanism ensures that autonomous driving functions prioritize safety, thereby more reliably achieving the driving goals of safety, comfort, and efficiency. Furthermore, this rulebook supports flexible expansion of rules or adjustment of rule priorities in different scenarios to meet the different decision-making focuses of each scenario.

[0090] The real-time scene information also includes real-time scene features, such as traffic density, weather conditions, and different driving modes (e.g., comfort mode, sport mode). For example, in heavy rain, the priority of "safety rules" is automatically increased.

[0091] By automatically optimizing trajectory decision-making logic based on rule book priorities using real-time scene information, when a new decision rule needs to be added, it can simply be inserted into the appropriate priority position in the rule book according to its importance, without manual intervention or modification or adjustment of any existing rules, thus reducing system maintenance costs and iteration cycles. At the same time, it can smooth out abrupt changes in rule priorities, avoiding discontinuities in decision logic caused by scene switching.

[0092] 303. Sort the rule information in descending order of priority.

[0093] For example, based on the priority of each rule information, the rule information is sorted in descending order of priority to obtain the sorting result, where the rule information with the highest priority is the first rule information and the rule information with the lowest priority is the last rule information.

[0094] 304. For any trajectory combination consisting of two initial trajectories, let i=1, take the i-th rule information as the reference rule information, compare the two initial trajectories in the trajectory combination, and determine the candidate trajectory in the trajectory combination based on the comparison result.

[0095] For example, all initial trajectories are paired to obtain at least one trajectory combination consisting of any two initial trajectories. Based on the sorted rule information, the first rule information is selected as the reference rule information in sequence. Using the judgment conditions in this reference rule information, the two initial trajectories in all trajectory combinations are compared pairwise to obtain the comparison result for each trajectory combination under each rule information. Based on the comparison results for each trajectory combination under all rule information, if a winning initial trajectory can be determined for each trajectory combination, then the winning initial trajectory is determined as a candidate trajectory for that trajectory combination. Then, the comparison of these two initial trajectories in that trajectory combination ends.

[0096] In one example, the rule information for the N trajectory decision dimensions includes safety rules, comfort rules, and efficiency rules. When the safety rule has the highest priority, the two initial trajectories in each trajectory combination are compared pairwise according to the safety rule. This includes determining whether trajectory A is safer than trajectory B; specifically, it can be determined whether trajectory A is farther from all obstacles than trajectory B. If the safety rule clearly determines that one of the initial trajectories is safer, then the safer initial trajectory is selected as the candidate trajectory, and the comparison between these two initial trajectories ends. Alternatively, when the comfort rule has the highest priority, the two initial trajectories in each trajectory combination are compared pairwise according to the comfort rule. This includes determining whether the driving comfort of trajectory A is higher than that of trajectory B. If the safety rule clearly determines that one of the initial trajectories has higher driving comfort, then the initial trajectory with higher driving comfort is selected as the candidate trajectory, and the comparison between these two initial trajectories ends.

[0097] In one possible implementation, step 304 includes the following steps: calculating the trajectory cost value corresponding to the two initial trajectories in the trajectory combination according to the reference rule information; if the difference between the trajectory cost values ​​of the two initial trajectories is greater than the comparison threshold, then the initial trajectory with the lower trajectory cost value in the trajectory combination is determined as the candidate trajectory of the trajectory combination.

[0098] Specifically, Figure 5 A flowchart illustrating a rulebook-based autonomous driving trajectory decision-making method provided in this application is shown below. Figure 5 As shown, the various trajectory decision dimensions to be considered are abstracted into independent rule information, used to compare the merits of two initial trajectories in their respective trajectory decision dimensions. Simultaneously, a corresponding cost function is designed for each rule information to calculate the trajectory cost value (hereinafter referred to as "cost value") for each initial trajectory under multiple rule information, quantifying the trajectory's performance under that rule information (e.g., minimum distance to obstacles, rate of curvature change, etc.). The rule information is sorted according to its priority to construct a rulebook. Each rule information in the rulebook is used sequentially as a reference rule information to compare the trajectory cost values ​​of multiple initial trajectories. A comparison threshold can be set; for each trajectory combination, if the difference between the trajectory cost values ​​of two initial trajectories under the reference rule information is greater than the comparison threshold, the initial trajectory with the lower cost value is considered superior, and thus the initial trajectory with the lower cost value in that trajectory combination is identified as a candidate trajectory for that combination.

[0099] In one example, N trajectory decision dimensions include safety, comfort, and efficiency. For the safety dimension, corresponding safety rules are configured, including determining whether trajectory A is farther from all obstacles than trajectory B. This means that the cost of trajectory A under the safety rule and the cost of trajectory B under the same safety rule can be obtained based on the distances between trajectory A and all obstacles, and between trajectory B and all obstacles, respectively. For the comfort dimension, corresponding comfort rules are configured, including determining whether the curvature / acceleration of trajectory A is smaller than that of trajectory B. This means that the cost of trajectory A under the comfort rule and the cost of trajectory B under the same comfort rule can be obtained based on the curvature / acceleration of trajectory A and trajectory B, respectively. For the efficiency dimension, corresponding efficiency rules are configured, including determining whether the estimated arrival time of trajectory A is shorter than that of trajectory B. This means that the cost of trajectory A under the efficiency rule and the cost of trajectory B under the same efficiency rule can be obtained based on the estimated arrival times of trajectory A and trajectory B, respectively.

[0100] In one example, Figure 6 A schematic diagram of a pairwise comparison process for a rulebook is provided as an embodiment of this application, such as... Figure 6As shown, using N rules (including safety rules, comfort rules, traffic rule compliance rules, and efficiency rules) from a pre-constructed rulebook, each trajectory combination is compared pairwise. The highest priority safety rule in the rulebook is selected, and this safety rule is used to compare the two initial trajectories in the current trajectory combination, including trajectory A and trajectory B. The cost value of trajectory A under this safety rule is calculated to be 0.1, and the cost value of trajectory B under this safety rule is calculated to be 0.8. When the safety rule can clearly determine that one of the trajectories is better, that is, the absolute value of the difference between 0.1 and 0.8 (0.7) is greater than the set comparison threshold (0.5), then the trajectory A with the lower cost value is better. The better trajectory A is immediately selected as the winner, and the comparison of the two initial trajectories ends.

[0101] By calculating the trajectory cost of two initial trajectories under the reference rule information and introducing a comparison threshold for pairwise comparison, the candidate trajectories under the reference rule information can be distinguished. This can intuitively reflect the entire trajectory comparison process and make the decision-making logic transparent.

[0102] 305. If the candidate trajectory in the trajectory combination cannot be determined based on the comparison results, let i = i + 1 and return to step 304.

[0103] For example, for each trajectory combination, when i=1, after comparing each trajectory combination pairwise based on the first rule information, if the candidate trajectory in the trajectory combination cannot be determined based on the comparison result, i.e., the rule information cannot distinguish the superiority or inferiority of the two initial trajectories in the trajectory combination, then let i=i+1, and use the second rule information as reference rule information for trajectory comparison. Specifically, according to the judgment condition in the second rule information, the two initial trajectories in the trajectory combination are compared pairwise to obtain the comparison result of the trajectory combination under the second rule information. If the winning initial trajectory in the trajectory combination can be determined based on the comparison result of the trajectory combination under the second rule information, then the winning initial trajectory is determined as a candidate trajectory, and the pairwise comparison of the trajectory combination ends. If the winning initial trajectory in the trajectory combination still cannot be determined based on the comparison result of the trajectory combination under the second rule information, then let i=i+1, and use the third rule information as reference rule information for trajectory comparison. That is, based on the judgment conditions in the third rule information, the two initial trajectories in the trajectory combination are compared pairwise, and so on, until the winning initial trajectory in the trajectory combination can be determined based on the comparison results of the trajectory combination under the i-th rule information.

[0104] In one example, Figure 7 Another schematic diagram of the pairwise comparison process of the rulebook provided in this application embodiment is shown below. Figure 7As shown, using N rules from the constructed rulebook, each trajectory combination is compared pairwise. The rule with the highest priority in the rulebook is extracted and used to compare the two current trajectories. If the rule cannot distinguish the superiority or inferiority of the two trajectories (i.e., the cost difference between the two trajectories under this rule is small, and the cost difference of 0.3 is less than the cost difference threshold of 0.5), then the rule with the next lower priority in the rulebook is extracted, and the comparison continues until a result can be obtained.

[0105] Through a hierarchical decision-making mechanism, high-priority rules determine the quality of trajectories first, while low-priority rules only take effect when high-priority rules cannot distinguish between them, thus avoiding conflicts between dimensions. Furthermore, the decision-making logic at each step is traceable, clearly indicating which rule determines the outcome of a trajectory at which priority level, thereby improving the transparency of trajectory decision-making.

[0106] In one possible implementation, after step 305, the method further includes: adjusting the comparison threshold when i=N and the candidate trajectory for the trajectory combination cannot be determined based on the comparison result; and returning to step 304.

[0107] Specifically, for each trajectory combination, when i=N, if after comparing each trajectory combination pairwise based on the Nth rule information, it is still impossible to determine the candidate trajectory for that trajectory combination based on the comparison results (i.e., no result can be determined after traversing all rule information), then the comparison threshold for distinguishing the quality of the trajectories is adjusted. For example, the comparison threshold is reduced, and a new round of comparison is performed, i.e., i=1, using the first rule information as the reference rule information, comparing the trajectory cost values ​​of the two initial trajectories. For each trajectory combination, if the difference between the trajectory cost values ​​of the two initial trajectories in the trajectory combination under the first rule information is greater than the adjusted comparison threshold, then the initial trajectory with the lower trajectory cost value is better, and the initial trajectory with the lower trajectory cost value in that trajectory combination is determined as the candidate trajectory for that trajectory combination. If the candidate trajectory in the trajectory combination cannot be determined based on the comparison result under the first rule information, then let i=i+1, take the second rule information as the reference rule information, and combine the trajectory cost of the two initial trajectories under the second rule information and the adjusted comparison threshold to compare the trajectories until a better initial trajectory can be found.

[0108] Furthermore, during the comparison, the number of times the threshold is adjusted is monitored in real time, i.e. the number of comparison rounds. If there is still no result after the number of comparison rounds is greater than the preset number (e.g., 10 times), one trajectory is randomly selected or the first trajectory is selected by default.

[0109] By introducing a "dynamic threshold adjustment" mechanism in pairwise trajectory comparisons, if trajectories still cannot be distinguished after more than a preset number of comparison rounds, the threshold is gradually lowered to expand the comparison range. The dynamic threshold can adapt to the magnitude of trajectory differences in different scenarios, avoiding misjudgments caused by fixed thresholds and improving decision-making accuracy; by gradually relaxing the threshold, the probability of distinguishing trajectories is increased, and the probability of random selection is reduced.

[0110] 306. Construct a directed graph based on at least one candidate trajectory and each initial trajectory; wherein each vertex in the directed graph is used to represent each initial trajectory, and each arrow in the arrow pointing table between two vertices points to the candidate trajectory between the corresponding two initial trajectories.

[0111] For example, combined Figure 8 The vehicle corresponds to four initial trajectories: trajectory A, trajectory B, trajectory C, and trajectory D. Based on the graph construction rules, each initial trajectory is represented as a vertex. For every two vertices, the arrow pointing to each other in the arrowhead table points to a candidate trajectory between the corresponding two initial trajectories, thus obtaining a directed graph. For example, for the trajectory combination of trajectory A and trajectory B, the comparison result is that B is better than A, and B is a candidate trajectory in this combination. Therefore, the arrow connecting trajectory A and trajectory B points from trajectory A to trajectory B. For the trajectory combination of trajectory A and trajectory B, the comparison result is that A is better than D, and A is a candidate trajectory in this combination. Therefore, the arrow connecting trajectory A and trajectory D points from trajectory D to trajectory A.

[0112] 307. Based on the in-degree of each vertex in the directed graph, determine the optimal trajectory of the vehicle from each initial trajectory.

[0113] Based on the constructed directed graph, for each vertex, the in-degree of each vertex is determined to represent the number of arrows pointing to that vertex in the directed graph. Based on preset rules, the in-degree of each vertex is analyzed to determine the optimal vertex, and the initial trajectory corresponding to the optimal vertex is determined as the optimal trajectory of the vehicle. Figure 8 A rulebook directed graph provided for embodiments of this application, such as Figure 8 Of the four initial trajectories, trajectory A has an in-degree of 1, trajectory B has an in-degree of 3, trajectory C has an in-degree of 2, and trajectory D has an in-degree of 0. Trajectory B has the largest in-degree, therefore, among the four initial trajectories, trajectory B is the optimal trajectory.

[0114] In one implementation, based on preset rules, the average in-degree of each vertex is calculated, and vertices with an in-degree greater than or equal to this average are identified as target vertices. The initial trajectory corresponding to the target vertex with the shortest expected arrival time is determined as the vehicle's optimal trajectory, based on the estimated arrival time of the initial trajectory of the initial trajectory of the target vertex. Alternatively, the initial trajectory corresponding to the target vertex with the shortest expected distance is determined as the vehicle's optimal trajectory. Or, the initial trajectory corresponding to the target vertex with both the shortest expected distance and the shortest expected arrival time is determined as the vehicle's optimal trajectory.

[0115] By constructing a directed graph, the comparison results between trajectories can be displayed intuitively, which facilitates problem localization (such as clarifying that a certain trajectory is eliminated due to low in-degree). Trajectory decisions are made based on in-degree (the number of arrows pointing to a better trajectory), reducing decision bias caused by single comparison errors and further improving the accuracy of trajectory decisions.

[0116] In one possible implementation, step 307 includes the following steps: determining the vertex with the largest in-degree, and determining the initial trajectory corresponding to the vertex with the largest in-degree as the optimal trajectory for the vehicle.

[0117] Specifically, such as Figure 8 As shown, for each vertex representing the initial trajectory, the vertex with the largest in-degree is retained, and all other vertices except the one with the largest in-degree are deleted. The number of vertices with the largest in-degree is determined. If the number is determined to be 1, the initial trajectory corresponding to the vertex with the largest in-degree is determined as the optimal trajectory for the vehicle.

[0118] In one example, Figure 9 This application provides a schematic diagram of a rulebook forming a loop, as shown in the embodiment of the present application. Figure 9 As shown, if it is determined that all vertices in the directed graph have the same in-degree, meaning the rules form a cycle and cannot be compared, the set comparison threshold is adjusted (e.g., lowered) to break the cycle. Then, based on the adjusted threshold, all initial trajectories are compared pairwise according to the priority of the rule information from high to low to obtain multiple candidate trajectories. A new directed graph is then reconstructed based on these candidate trajectories to determine the optimal trajectory. If the directed graph still forms a cycle after continuous iterations of constructing the directed graph equal to or greater than a preset number (e.g., 10 times), it indicates that the differences between the initial trajectories are too small. In this case, one initial trajectory is randomly selected, or the first one is selected by default, and it becomes the final optimal trajectory.

[0119] In one possible implementation, if the number of vertices with the largest in-degree is greater than 1, step 307 further includes: taking the initial trajectory corresponding to the vertex with the largest in-degree as the new initial trajectory, and returning to execute step 303.

[0120] Specifically, for each vertex representing an initial trajectory, the vertex with the highest in-degree is retained, and all other vertices except the one with the highest in-degree are deleted. The number of vertices with the highest in-degree is determined. If this number is greater than 1, the initial trajectories corresponding to all vertices with the highest in-degree are determined as new initial trajectories. For any two new initial trajectories, a new trajectory combination is formed. Let i=1, and use the first rule information as reference rule information to compare the two new initial trajectories in each trajectory combination. Based on the comparison result, the optimal trajectory for the candidate trajectories in each new trajectory combination is determined. For each new trajectory combination, if a candidate trajectory cannot be determined based on the comparison result, let i=i+1, and use the second rule information as reference trajectory rule information to compare the two initial trajectories in the new trajectory combination. Based on the comparison result, the candidate trajectory for the new trajectory combination is determined.

[0121] In this embodiment, based on the above embodiments, on the one hand, the priority order of rules can be determined simply by business logic and understanding, making the decision logic clear and intuitive, avoiding the "seesaw effect" between dimensions, and reducing the complexity of parameter tuning; on the other hand, by dynamically switching, adding or deleting rules, or adjusting rule priorities, the decision system can be easily adapted to the needs of different scenarios, improving the generalization performance of the system.

[0122] Figure 10 This is a schematic diagram of the structure of a trajectory determination device provided in an embodiment of this application, as shown below. Figure 10 As shown, the device includes:

[0123] The acquisition module 401 is used to acquire at least two initial trajectories of the vehicle;

[0124] The first determining module 402 is used to determine the rule information corresponding to each of the N trajectory decision dimensions; wherein the rule information has priority, and N is a positive integer greater than or equal to 2;

[0125] The comparison module 403 is used to determine reference rule information from each rule information according to the priority of the rule information, and to compare each initial trajectory pairwise according to the reference rule information to obtain at least one candidate trajectory.

[0126] The second determining module 404 is used to determine the optimal trajectory of the vehicle among each initial trajectory based on at least one candidate trajectory.

[0127] In one possible implementation, the comparison module 403 is specifically used for: sorting the rule information according to the priority of the rule information from high to low; for any trajectory combination composed of two initial trajectories, let i=1, use the i-th rule information as the reference rule information to compare the two initial trajectories in the trajectory combination, and determine the candidate trajectory in the trajectory combination according to the comparison result; if the candidate trajectory in the trajectory combination cannot be determined according to the comparison result, let i=i+1, and return to execute the step of using the i-th rule information as the reference trajectory rule information to compare the two initial trajectories in the trajectory combination, and determine the candidate trajectory in the trajectory combination according to the comparison result.

[0128] In one possible implementation, the comparison module 403 is specifically used to: calculate the trajectory cost value corresponding to the two initial trajectories in the trajectory combination according to the reference rule information; if the difference between the trajectory cost values ​​of the two initial trajectories is greater than the comparison threshold, then the initial trajectory with the lower trajectory cost value in the trajectory combination is determined as the candidate trajectory of the trajectory combination.

[0129] In one possible implementation, the comparison module 403 is further specifically used for: adjusting the comparison threshold when i=N and the candidate trajectory of the trajectory combination cannot be determined based on the comparison result; returning to the execution command i=1, using the i-th rule information as the reference trajectory rule information to compare the two initial trajectories in the trajectory combination, and determining the candidate trajectory in the trajectory combination based on the comparison result.

[0130] In one possible implementation, the second determining module 404 is specifically used to: construct a directed graph based on at least one candidate trajectory and each initial trajectory; wherein each vertex in the directed graph is used to represent each initial trajectory, and each arrow in the arrow pointing table between two vertices points to a candidate trajectory between the corresponding two initial trajectories; and determine the optimal trajectory of the vehicle from each initial trajectory based on the in-degree of each vertex in the directed graph.

[0131] In one possible implementation, the second determining module 404 is specifically used to: determine the vertex with the largest in-degree, and determine the initial trajectory corresponding to the vertex with the largest in-degree as the optimal trajectory of the vehicle.

[0132] In one possible implementation, when the number of vertices with the largest in-degree is greater than 1, the second determining module 404 is further specifically used to: take the initial trajectory corresponding to the vertex with the largest in-degree as the new initial trajectory, and return to execute the step of determining reference rule information from each rule information according to the priority of the rule information, and comparing each initial trajectory pairwise according to the reference rule information to obtain at least one candidate trajectory.

[0133] In one possible implementation, the first determining module 402 is specifically used to: determine the rule information corresponding to each trajectory decision dimension and / or the priority of each rule information based on real-time scene information.

[0134] The apparatus in this embodiment can execute the technical solutions in the above method. Its specific implementation process and technical principles are the same, and will not be repeated here.

[0135] Figure 11 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application, such as... Figure 11 As shown, the electronic device includes: a memory 501 and a processor 502; the memory 501 is a memory used to store instructions executable by the processor 502.

[0136] The processor 502 is configured to perform the method provided in the above embodiments.

[0137] The electronic device also includes a receiver 503 and a transmitter 504. The receiver 503 is used to receive instructions and data sent by other devices, and the transmitter 504 is used to send instructions and data to external devices.

[0138] The specific implementation process of the processor can be found in the above method embodiments, and its implementation principle and technical effect are similar, so it will not be repeated here.

[0139] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0140] This application provides a vehicle that includes the electronic equipment described in the above embodiments. The electronic equipment stores computer execution instructions, which, when executed on a computer, cause the computer to perform the technical solutions described in the above embodiments.

[0141] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed on a computer, cause the computer to perform the technical solutions described above.

[0142] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory, electrically erasable programmable read-only memory, erasable programmable read-only memory, programmable read-only memory, read-only memory, magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0143] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. The readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an application-specific integrated circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in a device.

[0144] This application also provides a computer program product, which includes a computer program stored in a computer-readable storage medium. At least one processor can read the computer program from the computer-readable storage medium, and when the at least one processor executes the computer program, it can implement the technical solutions in the above embodiments.

[0145] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as magnetic disks or optical disks.

[0146] Finally, it should be noted that other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein, and is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. A trajectory determination method, characterized in that, include: Obtain at least two initial trajectories for the vehicle; Determine the rule information corresponding to each of the N trajectory decision dimensions; wherein the rule information has a priority, and N is a positive integer greater than or equal to 2; Reference rule information is determined from each rule information according to the priority of the rule information, and each initial trajectory is compared pairwise according to the reference rule information to obtain at least one candidate trajectory. A directed graph is constructed based on the at least one candidate trajectory and each of the initial trajectories; wherein each vertex in the directed graph is used to represent each of the initial trajectories, and the arrow pointing between two vertices indicates that each arrow points to a candidate trajectory between the corresponding two initial trajectories; The optimal trajectory of the vehicle is determined from each of the initial trajectories based on the in-degree of each vertex of the directed graph.

2. The method according to claim 1, characterized in that, The step of determining reference rule information from each rule information according to the priority of the rule information, and comparing each initial trajectory pairwise according to the reference rule information to obtain at least one candidate trajectory, includes: The rule information is sorted in descending order of priority. For any trajectory combination consisting of two initial trajectories, let i=1, and use the i-th rule information as reference rule information to compare the two initial trajectories in the trajectory combination, and determine the candidate trajectory in the trajectory combination based on the comparison result. If the candidate trajectory in the trajectory combination cannot be determined based on the comparison result, then let i = i + 1, and return to the step of using the i-th rule information as reference rule information to compare the two initial trajectories in the trajectory combination, and determine the candidate trajectory in the trajectory combination based on the comparison result.

3. The method according to claim 2, characterized in that, The step of comparing two initial trajectories in the trajectory combination and determining candidate trajectories in the trajectory combination based on the comparison result includes: Based on the reference rule information, calculate the trajectory cost value corresponding to the two initial trajectories in the trajectory combination respectively; If the difference between the trajectory values ​​of two initial trajectories is greater than the comparison threshold, then the initial trajectory with the lower corresponding trajectory value in the trajectory combination is determined as the candidate trajectory of the trajectory combination.

4. The method according to claim 3, characterized in that, The method further includes: When i=N and the candidate trajectory of the trajectory combination cannot be determined based on the comparison results, the comparison threshold is adjusted. Return to the step of setting i=1, using the i-th rule information as reference rule information to compare the two initial trajectories in the trajectory combination, and determining the candidate trajectory in the trajectory combination based on the comparison result.

5. The method according to claim 1, characterized in that, The step of determining the optimal trajectory of the vehicle from each of the initial trajectories based on the in-degree of each vertex of the directed graph includes: The vertex with the largest in-degree is determined, and the initial trajectory corresponding to the vertex with the largest in-degree is determined as the optimal trajectory of the vehicle.

6. The method according to claim 5, characterized in that, When the number of vertices with the highest in-degree is greater than 1, the method further includes: The initial trajectory corresponding to the vertex with the largest in-degree is taken as the new initial trajectory, and the process is returned to execute the steps of determining reference rule information from each rule information according to the priority of the rule information, and comparing each initial trajectory pairwise according to the reference rule information to obtain at least one candidate trajectory.

7. The method according to any one of claims 1-6, characterized in that, The rule information corresponding to each of the N trajectory decision dimensions includes: Based on real-time scene information, determine the rule information corresponding to each trajectory decision dimension and / or the priority of each rule information.

8. An electronic device, characterized in that, include: Memory, processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the processor to perform the method as described in any one of claims 1-7.

9. A vehicle, characterized in that, The vehicle includes the electronic equipment as described in claim 8.

Images