Automatic driving decision-making method, device, computer equipment, readable storage medium and program product

Abstract

This application relates to an autonomous driving decision-making method, apparatus, computer device, computer-readable storage medium, and computer program product. The method comprises: obtaining environmental information corresponding to the location of a target object, state information of the target object, and target rules; analyzing violation risks and safety risks based on the target rules, environmental information, and state information, respectively, to construct a rule potential field and a safety potential field corresponding to the target object; constructing a risk model based on the rule potential field and the safety potential field, and determining a target control strategy based on the quantified safety and violation risk characteristics in the risk model. This method can improve the safety of autonomous driving.

Description

Technical Field

[0001] The present application relates to the field of autonomous driving technology, and in particular to an autonomous driving decision-making method, apparatus, computer equipment, computer-readable storage medium, and computer program product. Background Art

[0002] With the development of autonomous driving technology, human-driven vehicles will share the road with autonomous vehicles. At this time, autonomous vehicles will also need to follow traffic rules to ensure the safety of all vehicles.

[0003] In traditional technology, an autonomous vehicle determines the target rules it must currently follow based on the road section it is on and the traffic rules corresponding to that section. The target rules (for example, the target rules include the driving speed corresponding to the current section) are used as constraints, combined with safety and collision avoidance, as conditions for decision-making and planning to guide the autonomous vehicle to perform automated driving.

[0004] However, in traditional technologies, the decision-making of autonomous vehicles strictly adheres to target rules, resulting in poor safety of autonomous vehicles. Summary of the Invention

[0005] Based on this, it is necessary to provide an autonomous driving decision-making method, device, computer equipment, computer-readable storage medium and computer program product to address the above-mentioned technical problems.

[0006] In a first aspect, the present application provides an autonomous driving decision-making method, comprising:

[0007] Obtaining the environment information corresponding to the location of the target object, the state information of the target object, and the target rules;

[0008] Analyze the violation risk and security risk respectively according to the target rule, the environment information and the status information, and construct a rule potential field and a security potential field corresponding to the target object;

[0009] A risk model is constructed based on the rule potential field and the safety potential field, and a target control strategy is determined based on the quantified safety features and violation risk features in the risk model.

[0010] In one embodiment, the step of obtaining the environment information corresponding to the location of the target object, the state information of the target object, and the target rule includes:

[0011] Obtaining environmental information corresponding to the location of the target object, state information of the target object, and initial rules;

[0012] Priority ranking is performed on each of the initial rules according to the environmental information and the state information of the target object, and the target rule is obtained by screening according to the ranking result.

[0013] In one embodiment, prioritizing the initial rules according to the environment information and the state information of the target object, and filtering the target rules according to the ranking results, includes:

[0014] Determining a violation degree indicator based on the constraint indicator corresponding to each of the initial rules and the status information;

[0015] Determining a priority index of each of the initial rules according to the decision intention information of each of the initial rules and the violation degree index;

[0016] The initial rules are sorted according to the priority index to obtain a sorting result, and the target rule is screened and determined according to the sorting result.

[0017] In one embodiment, the target rules include speed constraints, position constraints and distance constraints, the rule potential field includes a first rule potential field corresponding to the speed risk model, a second rule potential field corresponding to the behavior risk model and a third rule potential field corresponding to the distance risk model, and the safety potential field is the target safety potential field corresponding to the distance risk model.

[0018] In one embodiment, analyzing the violation risk and the security risk respectively according to the target rule, the environment information, and the state information to construct the rule potential field and the security potential field corresponding to the target object includes:

[0019] Modeling the regular potential field of the speed risk model according to the speed constraint and the vehicle speed in the state information to obtain the first regular potential field;

[0020] Modeling the rule potential field of the behavior risk model according to the position constraint and the vehicle position in the state information to obtain the second rule potential field;

[0021] Modeling the regular potential field of the distance risk model according to the distance constraint, the own vehicle position, and each other vehicle position in the environmental information to obtain a third regular potential field;

[0022] The safety potential field of the distance risk model is modeled according to the preset safety distance, the state information and the environmental information to obtain the target safety potential field.

[0023] In one embodiment, the constructing of the risk model based on the rule potential field and the safety potential field includes:

[0024] Superimposing the third rule potential field and the target safety potential field to obtain a target distance risk model;

[0025] A risk model is constructed according to the speed risk model, the behavior risk model and the target distance risk model.

[0026] In one embodiment, determining a target control strategy based on the quantified security features and violation risk features in the risk model includes:

[0027] Post-processing the risk model to obtain a target risk model;

[0028] Based on the quantified safety features and violation risk features in the target risk model, trajectory planning is performed to obtain a target trajectory;

[0029] A target control strategy corresponding to the target object is determined based on the target trajectory and dynamic constraints.

[0030] In a second aspect, the present application also provides an autonomous driving decision-making device, comprising:

[0031] An acquisition module is used to acquire the environment information corresponding to the location of the target object, the state information of the target object, and the target rules;

[0032] A risk model construction module analyzes the violation risk and the security risk respectively according to the target rule, the environmental information and the status information, and constructs a rule potential field and a security potential field corresponding to the target object;

[0033] A determination module constructs a risk model based on the rule potential field and the safety potential field, and determines a target control strategy based on the quantified safety features and violation risk features in the risk model.

[0034] In one embodiment, the acquisition module is specifically used to acquire the environment information corresponding to the location of the target object, the state information of the target object and the initial rules;

[0035] Priority ranking is performed on each of the initial rules according to the environmental information and the state information of the target object, and the target rule is obtained by screening according to the ranking result.

[0036] In one embodiment, the acquisition module is specifically configured to determine a violation degree indicator based on the constraint indicator corresponding to each of the initial rules and the state information;

[0037] Determining a priority index of each of the initial rules according to the decision intention information of each of the initial rules and the violation degree index;

[0038] The initial rules are sorted according to the priority index to obtain a sorting result, and the target rule is screened and determined according to the sorting result.

[0039] In one embodiment, the target rules include speed constraints, position constraints and distance constraints, the rule potential field includes a first rule potential field corresponding to the speed risk model, a second rule potential field corresponding to the behavior risk model and a third rule potential field corresponding to the distance risk model, and the safety potential field is the target safety potential field corresponding to the distance risk model.

[0040] In one embodiment, the risk model building module is specifically configured to model the regular potential field of the speed risk model according to the speed constraint and the vehicle speed in the state information to obtain the first regular potential field;

[0041] Modeling the rule potential field of the behavior risk model according to the position constraint and the vehicle position in the state information to obtain the second rule potential field;

[0042] Modeling the regular potential field of the distance risk model according to the distance constraint, the own vehicle position, and each other vehicle position in the environmental information to obtain a third regular potential field;

[0043] The safety potential field of the distance risk model is modeled according to the preset safety distance, the state information and the environmental information to obtain the target safety potential field.

[0044] In one embodiment, the determining module is specifically configured to superimpose the third regular potential field and the target safety potential field to obtain a target distance risk model;

[0045] A risk model is constructed according to the speed risk model, the behavior risk model and the target distance risk model.

[0046] In one embodiment, the determination module is specifically configured to post-process the risk model to obtain a target risk model;

[0047] Based on the quantified safety features and violation risk features in the target risk model, trajectory planning is performed to obtain a target trajectory;

[0048] A target control strategy corresponding to the target object is determined based on the target trajectory and dynamic constraints.

[0049] In a third aspect, the present application further provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0050] Obtaining the environment information corresponding to the location of the target object, the state information of the target object, and the target rules;

[0051] Analyze the violation risk and security risk respectively according to the target rule, the environment information and the status information, and construct a rule potential field and a security potential field corresponding to the target object;

[0052] A risk model is constructed based on the rule potential field and the safety potential field, and a target control strategy is determined based on the quantified safety features and violation risk features in the risk model.

[0053] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:

[0054] Obtaining the environment information corresponding to the location of the target object, the state information of the target object, and the target rules;

[0055] Analyze the violation risk and security risk respectively according to the target rule, the environment information and the status information, and construct a rule potential field and a security potential field corresponding to the target object;

[0056] A risk model is constructed based on the rule potential field and the safety potential field, and a target control strategy is determined based on the quantified safety features and violation risk features in the risk model.

[0057] In a fifth aspect, the present application further provides a computer program product, comprising a computer program, which, when executed by a processor, implements the following steps:

[0058] Obtaining the environment information corresponding to the location of the target object, the state information of the target object, and the target rules;

[0059] Analyze the violation risk and security risk respectively according to the target rule, the environment information and the status information, and construct a rule potential field and a security potential field corresponding to the target object;

[0060] A risk model is constructed based on the rule potential field and the safety potential field, and a target control strategy is determined based on the quantified safety features and violation risk features in the risk model.

[0061] The aforementioned autonomous driving decision-making method, apparatus, computer device, computer-readable storage medium, and computer program product obtain environmental information corresponding to the target object's location, state information of the target object, and target rules; analyze violation risks and safety risks based on the target rules, environmental information, and state information, respectively, to construct a rule potential field and a safety potential field corresponding to the target object; construct a risk model based on the rule potential field and the safety potential field, and determine a target control strategy based on the safety and violation risk characteristics quantified in the risk model. This method constructs the rule potential field and the safety potential field corresponding to the target object using the target rules and environmental information at the current location, and superimposes the rule potential field and the safety potential field into a risk model. This method can quantify the safety and violation risks of the current environment in real time, enabling the autonomous vehicle to balance rule constraints and safety constraints, address scenarios where safety and rules conflict, and improve the safety of the autonomous vehicle. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following briefly introduces the drawings required for use in the embodiments of the present application or related technical descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying any creative work.

[0063] Figure 1 1 is a flow chart of an autonomous driving decision-making method according to an embodiment;

[0064] Figure 2 A schematic diagram of a process for determining target rules in one embodiment;

[0065] Figure 3 A schematic diagram of a process for prioritizing initial rules in one embodiment;

[0066] Figure 4 A schematic diagram of a process for constructing a regular potential field and a safety potential field in one embodiment;

[0067] Figure 5 A schematic diagram of a process for constructing a risk model in one embodiment;

[0068] Figure 6 FIG1 is a flow chart of post-processing steps for a risk model in one embodiment;

[0069] Figure 7 1 is a flowchart illustrating an example of an autonomous driving decision-making method according to an embodiment;

[0070] Figure 8 This is a structural block diagram of an autonomous driving decision-making device in one embodiment;

[0071] Figure 9 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0072] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0073] In one embodiment, Figure 1 As shown, a method for autonomous driving decision-making is provided. This embodiment uses the method applied to a terminal as an example. The terminal can be an onboard terminal of an autonomous driving vehicle. It is understood that the method can also be applied to a server, or to a system including a terminal and a server, and implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0074] Step 102: Obtain the environment information corresponding to the location of the target object, the state information of the target object, and the target rules.

[0075] In this embodiment, the target object is the autonomous driving vehicle itself. The autonomous driving vehicle includes an onboard terminal (referred to as the terminal in the following embodiments) that receives data from the perception system and makes autonomous driving decisions based on that data. In an autonomous driving system, the terminal collects information about the vehicle's current environment, its own status, and applicable rules, serving as the basis for subsequent risk analysis and decision-making.

[0076] Specifically, the perception system includes radar, cameras, and other onboard sensors to acquire environmental information and the vehicle's status. Environmental information includes other vehicles, traffic signs, and markings in the surrounding environment. Other vehicle information includes their position in the global coordinate system, their geometry, speed, and heading angle. Traffic sign and marking information includes their type, coordinates, and curvature. The vehicle's status information includes their position in the global coordinate system, their geometry, speed, and heading angle.

[0077] Target rules are the traffic rules corresponding to the vehicle's location. For example, on highways, target rules can be categorized into speed constraints, distance constraints, and behavioral constraints. Each type of rule may contain different requirements corresponding to different rules. The target rule is the most critical and immediate rule to be followed within each type of rule. It serves as the basis for subsequent autonomous driving decisions, ensuring that autonomous vehicles prioritize the target rules with the greatest safety impact and the most serious consequences of violations, thereby ensuring driving safety and compliance.

[0078] Step 104 : Analyze the violation risk and security risk respectively according to the target rules, environment information, and status information, and construct a rule potential field and a security potential field corresponding to the target object.

[0079] In this embodiment of the application, the terminal analyzes the potential violation risks (e.g., speeding, lane-blocking, etc.) and safety risks (e.g., the possibility of collision, loss of control, etc.) facing the vehicle based on the collected information (including target rules, environmental information, and status information). The results of this analysis are presented in the form of a "potential field," where the "rule potential field" quantifies the compliance risk of the vehicle's state relative to the rules, and the "safety potential field" reflects the safety risk between the vehicle and other vehicles.

[0080] Step 106 : constructing a risk model based on the rule potential field and the safety potential field, and determining a target control strategy based on the quantified safety features and violation risk features in the risk model.

[0081] In an embodiment of the present application, the terminal constructs a unified risk model using the rule potential field and the security potential field. The risk model comprehensively considers the behavior of the target object, environmental risks, and violation risks, and provides a comprehensive perspective. The rule potential field reflects the degree to which the target object complies with rules and regulations in the current environment, while the security potential field describes the potential dangers posed by the environment to the target object and its corresponding safety characteristics. The terminal quantifies the violation risk characteristics and safety characteristics based on the rule potential field and the security potential field, thereby identifying various possible risk scenarios. Based on the quantified potential field characteristics, the terminal can formulate precise target control strategies, including adjusting the actions of the target object, optimizing operating procedures, or implementing early warning measures to minimize violations and safety risks, and ensure the safety and compliance of the target object in the environment.

[0082] In the above-mentioned autonomous driving decision-making method, the rule potential field and safety potential field corresponding to the target object are constructed through the target rules and environmental information of the current location, and the rule potential field and safety potential field are superimposed into a risk model, which can quantify the safety risks and violation risks of the current environment in real time, so that the autonomous driving vehicle can balance the rule constraints and safety constraints, cope with scenarios where safety and rules conflict, and improve the safety of the autonomous driving vehicle.

[0083] In an exemplary embodiment, Figure 2 As shown, step 102 includes steps 202 to 204. Among them:

[0084] Step 202: Obtain the environment information corresponding to the location of the target object, the state information of the target object, and the initial rules.

[0085] In the embodiment of the present application, the environment information corresponding to the location of the target object includes lane information and other vehicle information in the environment. The lane information includes the lane line coordinates , Lane line type , lane id and lane width ; Other vehicle information including vehicle coordinates , vehicle speed , heading angle ,length and width And the vehicle's predicted trajectory information ,in, The state information of the target object is the state information of the vehicle itself, including ,in, is the vehicle coordinate, and are the longitudinal and lateral velocities of the vehicle, Represents the heading angle of the own vehicle, Characterizes the rate of change of the heading angle of the own vehicle over time.

[0086] Step 204 : Prioritize each initial rule according to the environment information and the state information of the target object, and select the target rule according to the ranking result.

[0087] In the embodiment of the present application, first, the terminal classifies the initial rules of the vehicle constraint mechanism based on different initial rules. This embodiment is explained by taking the highway scenario as an example. When the autonomous driving is driving on the highway, the main rules involved include speed limit, distance limit, lane change and overtaking rules. According to the different restrictions on vehicle driving behavior by different rules, the constraint mechanism of traffic rules on vehicles in the highway scenario can be divided into three categories of restrictions, including speed limit, distance limit and behavior restriction. Speed ​​limit rules belong to speed restrictions, distance limit rules belong to distance restrictions, lane change rules and overtaking rules include distance restrictions and behavior restrictions, among which the behavior restriction is that you cannot ride on the lane line for a long time when changing lanes.

[0088] The terminal then prioritizes each initial rule and prioritizes potentially conflicting rules within each type of initial rule. Specifically, after categorizing the initial rules, it is clear that the only speed-limiting rules on highways require no additional prioritization. The only behavioral restriction is not riding on lane markings for extended periods, which also requires no prioritization. Both distance limits and lane-changing overtaking rules require different distances between vehicles, so priorities are needed to determine the order in which vehicles follow them.

[0089] After all initial rules are prioritized, the terminal selects the target rules that the terminal needs to comply with first at the current moment based on the ranking results.

[0090] In this embodiment, the target rules that should be followed most at the current moment are screened out through the sorting results to ensure that the autonomous driving vehicle can prioritize the execution of the rules that have the greatest impact on safety and compliance in complex environments. Using the target rules as the basis for rule potential field modeling can improve the accuracy of the rule potential field and risk model, thereby improving the safety of autonomous driving.

[0091] In an exemplary embodiment, Figure 3 As shown, step 204 includes steps 302 to 306. Among them:

[0092] Step 302: Determine a violation degree index based on the constraint index and status information corresponding to each initial rule.

[0093] In the embodiment of the present application, the terminal is preset with a dynamic priority function that comprehensively evaluates the initial rules based on security requirements and violation degree indicators:

[0094]

[0095] in, Represents the security requirement indicator, Represents the weight coefficient of the security requirement index, Represents the violation level indicator, Represents the weight coefficient of the violation degree indicator, Indicates the priority index corresponding to each initial rule.

[0096] First, the terminal calculates the violation degree index according to the degree of violation of a certain rule by the vehicle. The greater the violation degree, the greater the impact on driving safety and surrounding vehicles, and the less conducive it is to a safe and efficient traffic environment. Therefore, rules with a large degree of violation should be given appropriate priority, that is, the violation degree index is quantified by calculating the degree to which the vehicle state quantity deviates from the rule compliance range.

[0097] Specifically, for distance limit rules, the rule stipulates that when the vehicle's speed is greater than 100 km / h and the vehicle ahead of it is the vehicle ahead of it, the compliance distance is 100 meters. When the vehicle's speed is less than 100 km / h and the vehicle ahead of it is the vehicle ahead of it, the compliance distance is 50 meters. The rule has a clear threshold. The violation level indicators corresponding to distance limit rules with clear thresholds are as follows:

[0098]

[0099] in, Indicates the compliance distance, Indicates the distance between the vehicle and the preceding vehicle. Representing the vehicle, Represents the preceding vehicle, Indicates the vehicle speed.

[0100] Regarding lane changing and overtaking rules, the rules stipulate that motor vehicles changing lanes must not affect the normal driving of related vehicles. However, the rules do not clearly define the threshold. Based on the rule digitization method, this embodiment digitizes the fuzzy threshold in the rule into a calculation expression. The corresponding violation degree index of the lane changing and overtaking rules containing the fuzzy threshold is as follows:

[0101]

[0102] in, Indicates the speed of his car. The TTC (Time-To-Collision) for longitudinal compliance is defined as 2.3 seconds in this embodiment.

[0103] Step 304 : determining the priority index of each initial rule based on the decision intention information and the violation degree index of each initial rule.

[0104] In the embodiments of the present application, safety requirements are a key factor in determining the priority of traffic rules. Specifically, the essential purpose of rules is to ensure driving safety. Therefore, when setting the priority of rules, the safety requirements behind the rules need to be considered. The terminal determines the safety requirement index based on the decision intent information of each initial rule. First, for rules that do not contain decision intent, such as speed limit rules (which stipulate the maximum speed of a vehicle, a clear and fixed value) and distance limit rules (which stipulate the minimum safe distance between vehicles, also a clear numerical requirement), since speed limit rules and distance limit rules are fundamental to ensuring driving safety and must be followed at all times, rules without decision intent are given a higher priority. For rules that contain decision intent, such as lane change rules and overtaking rules, since these rules involve complex decision intent, safety can be ensured by canceling or adjusting the decision. For example, if changing lanes could result in danger, the vehicle may choose not to change lanes. Because these rules involve complex decisions and safety can be ensured by not executing the decision, they have a relatively low priority. In other words, when they conflict with other rules, rules involving complex decision intent can be appropriately deferred. Therefore, the safety requirement index constructed in this embodiment is as follows:

[0105]

[0106] in Indicates the decision intention information involving rules, , Indicates the presence of decision intention.

[0107] Furthermore, the terminal determines the priority index of each initial rule according to the dynamic priority function, the security requirement index and the violation degree index.

[0108] Step 306: sort the initial rules according to the priority index to obtain a sorting result, and screen and determine the target rule according to the sorting result.

[0109] In an embodiment of the present application, the terminal uses a priority index to sort conflicting initial rules. The order of the sorting results reflects the rules that have the greatest impact on safety in the current scenario. These rules are used as the primary guiding constraints for autonomous driving decisions, while some rules with less impact on safety are flexibly discarded, allowing the terminal to select the target rules that are most suitable for the current scenario. Optionally, the sorting results can be stored in a priority queue, and the terminal dynamically adjusts the control strategy based on the target rules in the priority queue to cope with the ever-changing traffic environment and vehicle status.

[0110] In this embodiment, through the safety demand index, the terminal can prioritize the identification and processing of safety rules involving complex decisions and high risks, and prioritize compliance with basic rules to ensure that the vehicle meets basic safety requirements under any circumstances and avoid safety hazards caused by ignoring basic rules. Combining the safety demand index and the violation degree index can adapt to complex environments and quickly adjust the rule priority in the event of an emergency (such as emergency avoidance) to ensure the safety of autonomous driving.

[0111] In an exemplary embodiment, the target rules include speed constraints, position constraints and distance constraints, the rule potential field includes a first rule potential field corresponding to the speed risk model, a second rule potential field corresponding to the behavior risk model and a third rule potential field corresponding to the distance risk model, and the safety potential field is a target safety potential field corresponding to the distance risk model.

[0112] In an exemplary embodiment, Figure 4 As shown, step 104 includes steps 402 to 408. Among them:

[0113] Step 402 : Modeling the regular potential field of the speed risk model according to the speed constraint and the vehicle speed in the state information to obtain a first regular potential field.

[0114] In the embodiment of the present application, the terminal constructs the first rule potential field of the speed risk model by analyzing the vehicle speed in the vehicle status information and the speed constraint in the relevant speed limit rules. First, the terminal obtains the vehicle speed from the status information. and the upper and lower speed limits in the speed limit rules , build a speed risk model and control the vehicle speed within the compliance speed range. Among them, the first rule potential field is as follows:

[0115]

[0116] in, is the speed-limiting potential field, and are the strength coefficient and shape coefficient, respectively. represents the vehicle speed, Respectively represent the upper and lower speed limits of the current rule.

[0117] Step 404 : Model the rule potential field of the behavior risk model according to the position constraint and the vehicle position in the state information to obtain a second rule potential field.

[0118] In this embodiment of the present application, the terminal obtains the vehicle's position, traffic sign and marking information, and position constraints in relevant rules from the state information, and constructs a behavior risk model as a lane-related potential field to prevent the vehicle from crossing the lane boundary and preventing the vehicle from riding on the lane line for a long time. The second rule potential field is as follows:

[0119]

[0120]

[0121] in, represents the horizontal coordinate of the vehicle, Indicates lane width, represents the amplitude of the behavioral risk model, represents the total number of lanes, Indicates the The horizontal coordinate of the lane centerline of the lane, Represents the potential field peak of the dotted lane line on a multi-lane road, ranging from 0 to 1. In this embodiment, , and finally obtained This is the second rule potential field, which represents the risk state of the vehicle under lane constraints.

[0122] Step 406 : Model the regular potential field of the distance risk model according to the distance constraint, the position of the own vehicle, and the position of each other vehicle in the environmental information to obtain a third regular potential field.

[0123] In this embodiment of the present application, the distance risk model is constructed by superimposing the target safety potential field and the third rule potential field (a potential field representing the violation risk). For the third rule potential field, since the target rule can be a distance rule with a clear threshold or a distance rule with a fuzzy threshold, the terminal can construct a hierarchical violation risk potential field based on priority weighting, namely the third rule potential field. For rules with clear thresholds, such as the distance limit rule, the third rule potential field is constructed as follows:

[0124]

[0125]

[0126] in, and Respectively represent The strength and shape coefficient of the regular risk field of other cars, and Represent the vehicle and the The compliance distance between the other vehicle and the vehicle in the horizontal and longitudinal directions is: The compliance distance is 100 meters. When the speed of the vehicle is less than 100 km / h and the vehicle is the vehicle in front of the vehicle, the compliance distance is It's 50 meters. In this embodiment, it is set to 1.5.

[0127] For rules with fuzzy thresholds, such as lane-changing overtaking rules, this embodiment normalizes the fuzzy threshold into a computable expression by combining the violation time to collision (TTC) with the vehicle speed. The third rule potential field is constructed as follows:

[0128]

[0129] in, The TTC violation when the vehicle changes lanes to overtake is determined through statistical analysis of the data set. Indicates lane width.

[0130] Step 408 : Modeling the safety potential field of the distance risk model according to the preset safety distance, state information, and environmental information to obtain a target safety potential field.

[0131] In the embodiment of the present application, the target security risk model is constructed as an insurmountable potential field with an extremely high peak value but a small range to ensure safety. The terminal constructs the target security potential field as follows:

[0132]

[0133] in, and Respectively represent The strength and shape coefficient of the safety risk potential field of other vehicles, and Indicates the distance from the vehicle The longitudinal and lateral distances between the other vehicles, i.e. and , Indicates the longitudinal and lateral safety distances, Respectively represent the longitudinal position and lateral position of the other car, and Respectively represent The length and width of his car. Indicates the safe time interval between the vehicle and other vehicles. Indicates the The speed of his car, Represent the vehicle and the The longitudinal and lateral relative speeds of the other vehicles, Indicates the vehicle's comfortable acceleration.

[0134] In this embodiment, by constructing a speed risk model, a behavior risk model, and a distance risk model, the terminal comprehensively evaluates the driving risk of the vehicle from three aspects: speed constraints, behavior constraints, and distance constraints. This forms a multi-dimensional potential field that considers violation risks and safety risks. This will provide effective risk avoidance and decision-making guidance for the vehicle in complex traffic environments, ensuring the safety of autonomous driving in complex scenarios.

[0135] In an exemplary embodiment, Figure 5 As shown, step 106 includes steps 502 to 504. Among them:

[0136] Step 502: Superimpose the third rule potential field and the target safety potential field to obtain a target distance risk model.

[0137] In an embodiment of the present application, the terminal superimposes the third rule potential field and the target safety potential field through a priority index to form a target distance risk model, wherein the third rule potential field (graded violation risk potential field) is used to ensure that the vehicle complies with traffic rules during driving and avoids traffic conflicts and accidents caused by violations. It includes rules with clear thresholds and rules with fuzzy thresholds, and the rules with clear thresholds and the rules with fuzzy thresholds are hierarchically superimposed to comprehensively assess the vehicle's violation risk; the target safety potential field ensures that the vehicle can maintain a sufficient safety distance under any circumstances to avoid collisions or other dangerous situations. The target safety potential field involves calculating the minimum safety distance between vehicles to ensure that there is sufficient time and space for avoidance or braking in an emergency. The safety risk potential field is usually set to have an extremely high peak value within a very close range to force the vehicle to make an emergency response within a shorter safety distance. The safety risk potential field is a potential field with a clear threshold. Any object entering the area will be subjected to an extremely high "repulsive force" to ensure that the vehicle does not enter the dangerous area.

[0138] After determining the third rule potential field and the target safety potential field, the terminal performs inverse proportionality and normalization on the priority to use it as the weight, where:

[0139]

[0140] in, is the weight coefficient of the third rule potential field (violation risk model), , representing the distance limit rule and lane changing overtaking rule respectively.

[0141] By superimposing the target safety potential field and the third rule potential field (the priority-based hierarchical violation risk potential field), the target distance risk model is constructed as follows:

[0142]

[0143] Step 504: construct a risk model based on the speed risk model, the behavior risk model, and the target distance risk model.

[0144] In the embodiment of the present application, the terminal constructs a risk model by combining the speed risk model, the behavior risk model, and the distance risk model. The constructed risk model is as follows:

[0145]

[0146] In this embodiment, the third rule potential field (the hierarchical violation risk field) and the target safety potential field are inversely calculated and normalized based on priority indicators to determine their respective weights. These are then superimposed to form a target distance risk model. This effectively combines the violation risk assessment required to ensure vehicle compliance with traffic rules with the emergency response mechanism required to maintain a safe distance. Furthermore, by combining the speed risk model with the behavioral risk model, a comprehensive risk model is constructed, enabling a comprehensive assessment of multi-dimensional risks during vehicle operation. This not only enables precise control of vehicle safety and compliance, but also enables timely and appropriate decision-making in complex traffic environments, thereby enhancing the overall safety and reliability of the autonomous driving system.

[0147] In an exemplary embodiment, Figure 6 As shown, step 106 includes steps 602 to 606. Among them:

[0148] Step 602: Post-process the risk model to obtain a target risk model.

[0149] In an embodiment of the present application, the terminal post-processes the risk model through convex transformation and Taylor approximation. Specifically, the terminal converts the risk model into a quadratic convex function through convex transformation and Taylor approximation, so that the risk model can be applied to subsequent optimization problems for trajectory planning. Since the third rule potential field, the third rule potential field, the third rule potential field, and the target safety potential field in the current risk model are nonlinear and non-convex, this embodiment converts the potential field solution problem into a quadratic convex problem to improve its solution time efficiency.

[0150] In this embodiment, a risk potential field is defined in the direction of signed distance (SD) to generate repulsion. Therefore, the potential field of the risk model ( ) can be converted to function instead of ,in is the distance in the SD direction. In other words, for a potential field PF, , there is a function , making .

[0151] Take the target security risk potential field as an example to illustrate, assuming:

[0152]

[0153]

[0154] but:

[0155]

[0156] Subsequently, based on the second Taylor approximation, the original potential field function is approximated offline. The quadratic function is a compact convex quadratic approximation of the original PF function about the nominal point. First, the derivative of the PF function with respect to the global coordinate variable is calculated as follows:

[0157]

[0158] From this we can get the global coordinates The first-order gradient of:

[0159]

[0160] Similarly, the Hessian matrix of the PF function with respect to the global coordinate variables can also be calculated as:

[0161]

[0162] Finally, represents the Taylor operation point, then the PF function can be approximated as:

[0163]

[0164] in:

[0165]

[0166] Step 604 : performing trajectory planning based on the quantified safety features and violation risk features in the target risk model to obtain a target trajectory.

[0167] In an embodiment of the present application, a terminal constructs a Model Predictive Control (MPC) based on a target risk model to generate a target trajectory that satisfies vehicle dynamics constraints, thereby achieving compliant autonomous driving trajectory planning. MPC uses a vehicle dynamics model to predict future outputs using a preset initial state and a finite time step. It then uses an optimization algorithm to solve for the optimal control input sequence, taking into account the safety and violation risk characteristics quantified in various dynamic and target risk models, to minimize the deviation between the predicted output and the target trajectory. Subsequently, the control input is applied and updated, and corrections are made based on real-time feedback information to ensure that the system meets the constraints of the target risk model throughout the entire prediction and control process, achieving accurate target trajectory generation.

[0168] Step 606 : Determine a target control strategy corresponding to the target object based on the target trajectory and the dynamic constraints.

[0169] In the embodiment of the present application, the target control strategy is the maximum steering angle, the maximum longitudinal force and the maximum steering angle change. The terminal takes the vehicle state and control quantity as and , control system output ,in:

[0170]

[0171] in, The differential equation representing the state quantity, is the state transition matrix, is the control input matrix.

[0172] and

[0173]

[0174] in, and are the longitudinal and lateral positions of the vehicle, and are the longitudinal and lateral velocities of the vehicle, is the heading angle of the vehicle, and are the cornering stiffness of the front and rear tires, and are the distances from the vehicle's center of mass to the front and rear axles, is the rotational inertia of the vehicle, is the mass of the vehicle, is the longitudinal force of the vehicle, is the front wheel turning angle of the vehicle.

[0175] The cost function of MPC includes the potential field, tracking error and control amount, which is expressed as follows:

[0176]

[0177] The last two terms of the above equation can be written as:

[0178]

[0179]

[0180]

[0181] in, is the maximum steering angle, is the maximum longitudinal force, is the maximum steering angle change. , is the symbol for minimizing the cost function, is the discrete time step, and represent the prediction time domain and the control time domain respectively, Indicates the current moment, represents the control increment and control amount, Q is the state weight matrix, represents the control weight matrix.

[0182] In a specific embodiment, an autonomous driving decision-making method is provided, such as Figure 7 Shown, including:

[0183] Step 701: Obtaining the environment information corresponding to the location of the target object, the state information of the target object, and the highway rules;

[0184] Step 702: classify highway rules based on their constraints on vehicles.

[0185] Step 703, prioritize highway rules;

[0186] Step 704: construct a risk model based on the prioritized highway rules;

[0187] Step 705 , post-processing the risk model through convex transformation and linearization to convert the risk model into a linear quadratic convex function form;

[0188] Step 706: perform trajectory planning based on the post-processed risk model.

[0189] In this embodiment, the efficiency of the model solution is significantly improved by converting the complex and nonlinear risk model into a quadratic convex function and then performing offline approximation using the Taylor approximation method. Next, the quadratic approximate risk model is combined with a model predictive controller to generate a target trajectory that satisfies vehicle dynamics constraints. This approach comprehensively considers both safety characteristics and violation risk characteristics, achieving precise autonomous driving trajectory planning. Finally, an optimization algorithm is used to solve for the optimal control input sequence, ensuring that the system meets the constraints of the comprehensive risk model under real-time feedback correction, thereby improving the overall safety of the autonomous driving system.

[0190] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0191] Based on the same inventive concept, the present application also provides an autonomous driving decision-making device for implementing the aforementioned autonomous driving decision-making method. The solution provided by this device is similar to the solution described in the aforementioned method. Therefore, the specific limitations in one or more of the following autonomous driving decision-making device embodiments can be found in the aforementioned limitations on the autonomous driving decision-making method and will not be further elaborated here.

[0192] In an exemplary embodiment, Figure 8 As shown, an autonomous driving decision-making device 800 is provided, comprising: an acquisition module 801, a risk model construction module 802, and a determination module 803, wherein:

[0193] An acquisition module 801 is used to acquire the environment information corresponding to the location of the target object, the state information of the target object, and the target rules;

[0194] The risk model construction module 802 analyzes the violation risk and security risk respectively according to the target rules, environment information and status information, and constructs the rule potential field and security potential field corresponding to the target object;

[0195] The determination module 803 constructs a risk model based on the rule potential field and the safety potential field, and determines a target control strategy based on the quantified safety features and violation risk features in the risk model.

[0196] In one embodiment, the acquisition module 801 is specifically used to acquire the environment information corresponding to the location of the target object, the state information of the target object, and the initial rules;

[0197] The initial rules are prioritized based on the environment information and the status information of the target object, and the target rules are obtained based on the sorting results.

[0198] In one embodiment, the acquisition module 801 is specifically configured to determine a violation degree indicator based on the constraint indicator and status information corresponding to each initial rule;

[0199] Determine the priority index of each initial rule based on the decision intention information and violation degree index of each initial rule;

[0200] The initial rules are sorted according to the priority index to obtain the sorting results, and the target rules are screened and determined based on the sorting results.

[0201] In one embodiment, the target rules include speed constraints, position constraints and distance constraints, the rule potential field includes a first rule potential field corresponding to the speed risk model, a second rule potential field corresponding to the behavior risk model and a third rule potential field corresponding to the distance risk model, and the safety potential field is a target safety potential field corresponding to the distance risk model.

[0202] In one embodiment, the risk model building module 802 is specifically configured to model a regular potential field of the speed risk model according to the speed constraint and the vehicle speed in the state information to obtain a first regular potential field;

[0203] The rule potential field of the behavior risk model is modeled according to the position constraint and the position of the ego vehicle in the state information to obtain a second rule potential field;

[0204] The rule potential field of the distance risk model is modeled according to the distance constraint, the position of the own vehicle, and the position of each other vehicle in the environment information to obtain the third rule potential field;

[0205] The safety potential field of the distance risk model is modeled according to the preset safety distance, state information and environmental information to obtain the target safety potential field.

[0206] In one embodiment, the determination module 803 is specifically configured to superimpose the third rule potential field and the target safety potential field to obtain a target distance risk model;

[0207] A risk model is constructed based on the speed risk model, behavior risk model and target distance risk model.

[0208] In one embodiment, the determination module 803 is specifically used to post-process the risk model to obtain a target risk model;

[0209] Based on the quantified safety features and violation risk features in the target risk model, trajectory planning is performed to obtain the target trajectory;

[0210] The target control strategy corresponding to the target object is determined based on the target trajectory and dynamic constraints.

[0211] Each module in the aforementioned autonomous driving decision-making device may be implemented in whole or in part through software, hardware, or a combination thereof. Each module may be embedded in or independent of a processor in a computer device in hardware form, or may be stored in a computer device memory in software form, so that the processor can call and execute the corresponding operations of each module.

[0212] In an exemplary embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as shown in FIG. Figure 9As shown. The computer device includes a processor, memory, an input / output interface, a communication interface, a display unit, and an input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are connected to the system bus via the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals via wired or wireless means, and the wireless means can be implemented via Wi-Fi, a mobile cellular network, near-field communication (NFC), or other technologies. When executed by the processor, the computer program implements an autonomous driving decision-making method. The display unit of the computer device is used to form a visually visible image, and can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device casing, or an external keyboard, touchpad or mouse.

[0213] Those skilled in the art will understand that Figure 9 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0214] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:

[0215] Obtaining the environment information corresponding to the location of the target object, the state information of the target object, and the target rules;

[0216] Analyze the violation risk and security risk respectively according to the target rule, the environment information and the status information, and construct a rule potential field and a security potential field corresponding to the target object;

[0217] A risk model is constructed based on the rule potential field and the safety potential field, and a target control strategy is determined based on the quantified safety features and violation risk features in the risk model.

[0218] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0219] Obtaining environmental information corresponding to the location of the target object, state information of the target object, and initial rules;

[0220] Priority ranking is performed on each of the initial rules according to the environmental information and the state information of the target object, and the target rule is obtained by screening according to the ranking result.

[0221] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0222] Determining a violation degree indicator based on the constraint indicator corresponding to each of the initial rules and the status information;

[0223] Determining a priority index of each of the initial rules according to the decision intention information of each of the initial rules and the violation degree index;

[0224] The initial rules are sorted according to the priority index to obtain a sorting result, and the target rule is screened and determined according to the sorting result.

[0225] In one embodiment, the target rules include speed constraints, position constraints and distance constraints, the rule potential field includes a first rule potential field corresponding to the speed risk model, a second rule potential field corresponding to the behavior risk model and a third rule potential field corresponding to the distance risk model, and the safety potential field is the target safety potential field corresponding to the distance risk model.

[0226] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0227] Modeling the regular potential field of the speed risk model according to the speed constraint and the vehicle speed in the state information to obtain the first regular potential field;

[0228] Modeling the rule potential field of the behavior risk model according to the position constraint and the vehicle position in the state information to obtain the second rule potential field;

[0229] Modeling the regular potential field of the distance risk model according to the distance constraint, the own vehicle position, and each other vehicle position in the environmental information to obtain a third regular potential field;

[0230] The safety potential field of the distance risk model is modeled according to the preset safety distance, the state information and the environmental information to obtain the target safety potential field.

[0231] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0232] Superimposing the third rule potential field and the target safety potential field to obtain a target distance risk model;

[0233] A risk model is constructed according to the speed risk model, the behavior risk model and the target distance risk model.

[0234] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0235] Post-processing the risk model to obtain a target risk model;

[0236] Based on the quantified safety features and violation risk features in the target risk model, trajectory planning is performed to obtain a target trajectory;

[0237] A target control strategy corresponding to the target object is determined based on the target trajectory and dynamic constraints.

[0238] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0239] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0240] 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, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0241] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. In particular, any reference to memory, database, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The databases involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the various embodiments provided herein may be, but are not limited to, general-purpose processors, central processing units (CPUs), graphics processing units (GPUs), digital signal processors (DSPs), programmable logic devices (PLDs), quantum computing-based data processing logic devices, artificial intelligence (AI) processors, and the like.

[0242] The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0243] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. An automatic driving decision-making method, characterized in that: The method comprises: Obtaining the environment information corresponding to the location of the target object, the state information of the target object, and the target rules; Analyze the violation risk and security risk respectively according to the target rule, the environment information and the status information, and construct a rule potential field and a security potential field corresponding to the target object; Constructing a risk model based on the rule potential field and the safety potential field, and determining a target control strategy based on the quantified safety features and violation risk features in the risk model; The target rules include speed constraints, position constraints, and distance constraints. The rule potential field includes a first rule potential field corresponding to a speed risk model, a second rule potential field corresponding to a behavior risk model, and a third rule potential field corresponding to a distance risk model. The safety potential field is a target safety potential field corresponding to the distance risk model. The analyzing the violation risk and the security risk respectively according to the target rule, the environment information and the state information to construct the rule potential field and the security potential field corresponding to the target object includes: Modeling the regular potential field of the speed risk model according to the speed constraint and the vehicle speed in the state information to obtain the first regular potential field; Modeling the rule potential field of the behavior risk model according to the position constraint and the vehicle position in the state information to obtain the second rule potential field; Modeling the regular potential field of the distance risk model according to the distance constraint, the own vehicle position, and each other vehicle position in the environmental information to obtain a third regular potential field; Modeling the safety potential field of the distance risk model according to the preset safety distance, the state information and the environmental information to obtain the target safety potential field; The constructing of the risk model based on the rule potential field and the safety potential field includes: Superimposing the third rule potential field and the target safety potential field to obtain a target distance risk model; A risk model is constructed according to the speed risk model, the behavior risk model and the target distance risk model.

2. The method according to claim 1, characterized in that The acquiring of the environment information corresponding to the location of the target object, the state information of the target object, and the target rule includes: Obtaining environmental information corresponding to the location of the target object, state information of the target object, and initial rules; Priority ranking is performed on each of the initial rules according to the environmental information and the state information of the target object, and the target rule is obtained by screening according to the ranking result.

3. The method according to claim 2, characterized in that The prioritizing of the initial rules according to the environment information and the state information of the target object, and obtaining the target rule according to the prioritization results, includes: Determining a violation degree indicator based on the constraint indicator corresponding to each of the initial rules and the status information; Determining a priority index of each of the initial rules according to the decision intention information of each of the initial rules and the violation degree index; The initial rules are sorted according to the priority index to obtain a sorting result, and the target rule is screened and determined according to the sorting result.

4. The method according to claim 1, wherein Determining a target control strategy based on the quantified safety features and violation risk features in the risk model includes: Post-processing the risk model to obtain a target risk model; Based on the quantified safety features and violation risk features in the target risk model, trajectory planning is performed to obtain a target trajectory; A target control strategy corresponding to the target object is determined based on the target trajectory and dynamic constraints.

5. An automatic driving decision-making device, characterized in that: The device comprises: An acquisition module is used to acquire the environment information corresponding to the location of the target object, the state information of the target object, and the target rules; A risk model construction module analyzes the violation risk and the security risk respectively according to the target rule, the environmental information and the status information, and constructs a rule potential field and a security potential field corresponding to the target object; a determination module, constructing a risk model based on the rule potential field and the safety potential field, and determining a target control strategy based on the quantified safety features and violation risk features in the risk model; The target rules include speed constraints, position constraints, and distance constraints. The rule potential field includes a first rule potential field corresponding to a speed risk model, a second rule potential field corresponding to a behavior risk model, and a third rule potential field corresponding to a distance risk model. The safety potential field is a target safety potential field corresponding to the distance risk model. The risk model building module is specifically configured to model the regular potential field of the speed risk model according to the speed constraint and the vehicle speed in the state information to obtain the first regular potential field; Modeling the rule potential field of the behavior risk model according to the position constraint and the vehicle position in the state information to obtain the second rule potential field; Modeling the regular potential field of the distance risk model according to the distance constraint, the own vehicle position, and each other vehicle position in the environmental information to obtain a third regular potential field; Modeling the safety potential field of the distance risk model according to the preset safety distance, the state information and the environmental information to obtain the target safety potential field; The determination module is specifically configured to superimpose the third rule potential field and the target safety potential field to obtain a target distance risk model; A risk model is constructed according to the speed risk model, the behavior risk model and the target distance risk model.

6. The device according to claim 5, characterized in that The acquisition module is specifically used to acquire the environment information corresponding to the location of the target object, the state information of the target object and the initial rules; Priority ranking is performed on each of the initial rules according to the environmental information and the state information of the target object, and the target rule is obtained by screening according to the ranking result.

7. The device according to claim 6, characterized in that The acquisition module is specifically configured to determine a violation degree indicator based on the constraint indicator corresponding to each of the initial rules and the status information; Determining a priority index of each of the initial rules according to the decision intention information of each of the initial rules and the violation degree index; The initial rules are sorted according to the priority index to obtain a sorting result, and the target rule is screened and determined according to the sorting result.

8. The device according to claim 5, characterized in that The determination module is specifically used to post-process the risk model to obtain a target risk model; Based on the quantified safety features and violation risk features in the target risk model, trajectory planning is performed to obtain a target trajectory; A target control strategy corresponding to the target object is determined based on the target trajectory and dynamic constraints.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 4 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

Images