Vehicle automatic driving decision method, device, vehicle terminal and storage medium

By employing lateral decision-making strategies for roads and intersections in autonomous driving, and utilizing Markov reinforcement learning and finite state machines, the problem of high decision complexity in existing technologies is solved, thereby improving decision-making efficiency and accuracy.

CN114644016BActive Publication Date: 2024-11-29CHINA AUTOMOTIVE INNOVATION CORP
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
CN202210393556.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-04-14
Publication Date
2024-11-29
Estimated Expiration
2042-04-14

AI Technical Summary

Technical Problem

Existing vehicle decision-making modules employ complex decision-making methods that consume significant computational resources, resulting in low algorithm performance and decision-making efficiency.

Method used

We employ lateral decision-making strategies for roads and intersections, respectively, to make decisions in road driving scenarios and intersection driving scenarios. We utilize Markov reinforcement learning models and finite state machine decision-making strategies to reduce the complexity of the decision-making system.

Benefits of technology

It improves the efficiency and accuracy of autonomous driving decisions and reduces the consumption of computing resources.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114644016B_ABST
    Figure CN114644016B_ABST
Patent Text Reader

Abstract

The present application discloses a vehicle automatic driving decision method, device, vehicle-mounted terminal and storage medium, which obtains the current position information, current navigation information, current driving condition information and current road environment information of the road on which the target vehicle is traveling by obtaining the current position information of the target vehicle, the current navigation information, the current driving condition information and the current road environment information of the road on which the target vehicle is traveling; when the distance between the current position information of the target vehicle and the target intersection is greater than or equal to the preset distance threshold, the road lateral decision strategy is obtained, and the first lateral decision result of the target vehicle is obtained according to the road lateral decision strategy; otherwise, the intersection lateral decision strategy is obtained, and the second lateral decision result of the target vehicle is obtained according to the intersection lateral decision strategy; the driving trajectory of the target vehicle is determined according to the first lateral decision result and the second lateral decision result. The present application solves the problem of the complexity of the automatic driving decision model by adopting different decision methods in road conditions and intersection conditions, reduces the complexity of the decision model under road conditions, and improves the decision efficiency and accuracy.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of vehicle autonomous driving technology, and in particular to a vehicle autonomous driving decision-making method, device, vehicle-mounted terminal and storage medium. Background Art

[0002] An autonomous driving vehicle refers to an intelligent vehicle that senses the road environment through an on-board sensor system, automatically plans the driving route and controls the vehicle to reach the predetermined destination. The autonomous driving system is usually divided into map positioning, perception prediction, path navigation, decision planning, vehicle control and other parts. Among them, the decision planning module can include lateral decision-making and longitudinal decision-making.

[0003] Existing vehicle decision modules usually extract multiple factors that affect vehicle driving behavior, such as road line type, speed of surrounding obstacles and distance from the vehicle, traffic signs, predicted trajectory of obstacles, etc., and set quantitative indicators of related factors, and determine the conditions for decision state transfer based on the quantitative indicators. However, this decision-making method is difficult, consumes a lot of computing resources, and has low algorithm performance, resulting in low decision efficiency. Summary of the invention

[0004] In order to solve the technical problems that existing decision-making methods are difficult, consume a lot of computing resources, and have low algorithm performance and decision efficiency, the present invention provides a vehicle automatic driving decision-making method, device, vehicle-mounted terminal and storage medium, which adopts a road lateral decision strategy in road conditions and an intersection lateral decision strategy in intersection conditions, thereby reducing the complexity of the decision system model and improving the algorithm performance and decision efficiency.

[0005] On the one hand, an embodiment of the present application provides a vehicle automatic driving decision method, including:

[0006] Acquiring current location information, current navigation information, current driving condition information of a target vehicle and current road environment information of a road on which the target vehicle is traveling;

[0007] Based on the current position information and current road environment information of the target vehicle, determining the driving scene of the target vehicle, wherein the driving scene includes a road driving scene and an intersection driving scene;

[0008] When the driving scene of the target vehicle is the road driving scene, the current driving condition information and the current road environment information of the target vehicle are processed according to the road lateral decision strategy to obtain a first lateral decision result of the target vehicle;

[0009] When the driving scene of the target vehicle is the intersection driving scene, the current navigation information and the current road environment information of the target vehicle are processed according to the intersection lateral decision strategy to obtain a second lateral decision result of the target vehicle;

[0010] The driving trajectory of the target vehicle is determined according to the first lateral decision result and / or the second lateral decision result.

[0011] Further, the determining the driving scene of the target vehicle based on the current position information of the target vehicle and the current road environment information includes:

[0012] The current road environment information includes information of a front target intersection and a rear target intersection of the target vehicle;

[0013] When the distance between the current position information of the target vehicle and the front target intersection is greater than or equal to a first distance threshold or the distance between the current position information of the target vehicle and the rear target intersection is greater than or equal to a second distance threshold, the driving scene of the target vehicle is determined to be the road driving scene.

[0014] Further,

[0015] The determining of the driving scene of the target vehicle based on the current position information and the current road environment information of the target vehicle includes:

[0016] The current road environment information includes information of a front target intersection and a rear target intersection of the target vehicle;

[0017] When the distance between the current position information of the target vehicle and the front target intersection is less than a first distance threshold or the distance between the target position information and the rear target intersection is less than a second distance threshold, the driving scene of the target vehicle is determined to be the intersection driving scene.

[0018] Further, the acquiring of the road lateral decision strategy, processing the current driving condition information and the current road environment information of the target vehicle according to the road lateral decision strategy to obtain a first lateral decision result of the target vehicle, includes:

[0019] Receiving the road lateral decision model sent by the server;

[0020] Using the received road lateral decision model as the road lateral decision strategy;

[0021] Processing the current driving condition information and the current road environment information of the target vehicle according to the road lateral decision model to obtain the first lateral decision result of the target vehicle;

[0022] Among them, the road lateral decision model is obtained by training a preset model based on sample data of a sample vehicle. The preset model is a model based on Markov reinforcement learning. The sample data includes sample driving condition information of the sample vehicle and sample road environment information of the road on which the sample vehicle is traveling.

[0023] Furthermore, the method further comprises:

[0024] Receiving the updated road lateral decision model sent by the server;

[0025] Using the updated road lateral decision model as the road lateral decision strategy;

[0026] Among them, the updated road lateral decision model is obtained by training the preset model based on the updated sample data; the updated sample data is obtained by updating the sample data based on the current driving condition information of the target vehicle and the current road environment information.

[0027] Furthermore, the intersection lateral decision strategy is a lateral decision strategy based on a finite state machine, and the method further includes:

[0028] Obtaining preset navigation information of a preset vehicle, preset road environment information of a road on which the preset vehicle is traveling, a preset initial lateral decision result, and a preset target lateral decision result;

[0029] Establishing a mapping relationship between the preset navigation information, the preset road environment information, the preset initial lateral decision result and the preset target lateral decision result;

[0030] The mapping relationship is used as the lateral decision-making strategy for the intersection.

[0031] Further, the processing of the current navigation information and the current road environment information of the target vehicle according to the intersection lateral decision strategy to obtain a second lateral decision result of the target vehicle includes:

[0032] Obtaining a current initial lateral decision result of the target vehicle;

[0033] According to the mapping relationship, obtaining a preset lateral decision result corresponding to the current navigation information, the current road environment information and the current initial lateral decision result;

[0034] The preset lateral decision result corresponding to the current navigation information, the current road environment information and the current initial lateral decision result is used as the second lateral decision result.

[0035] Further, determining the driving trajectory of the target vehicle according to the first lateral decision result and / or the second lateral decision result includes:

[0036] Using the first lateral decision result and / or the second lateral decision result as a target lateral decision result of the target vehicle;

[0037] Determine the target vertical decision result based on the target horizontal decision result;

[0038] The driving track of the target vehicle is determined according to the target lateral decision result and the target longitudinal decision result.

[0039] In another aspect, the present invention provides a vehicle automatic driving decision-making device, comprising:

[0040] An acquisition module, used to acquire the current position information, current navigation information, current driving condition information of the target vehicle and the current road environment information of the road on which the target vehicle is traveling;

[0041] A scene determination module, used to determine the driving scene of the target vehicle based on the current position information of the target vehicle and the current road environment information, wherein the driving scene includes a road driving scene and an intersection driving scene;

[0042] A first lateral decision module is used to process the current driving condition information and the current road environment information of the target vehicle according to the road lateral decision strategy when the driving scene of the target vehicle is the road driving scene, so as to obtain a first lateral decision result of the target vehicle;

[0043] A second lateral decision module is used to process the current navigation information and the current road environment information of the target vehicle according to the intersection lateral decision strategy when the driving scene of the target vehicle is the intersection driving scene, so as to obtain a second lateral decision result of the target vehicle;

[0044] The trajectory determination module is used to determine the driving trajectory of the target vehicle according to the first lateral decision result and / or the second lateral decision result.

[0045] On the other hand, a computer device is provided, including a processor and a memory, wherein the memory stores at least one instruction or at least one program, and the at least one instruction or the at least one program is loaded and executed by the processor to implement the above-mentioned vehicle automatic driving decision-making method.

[0046] On the other hand, a computer-readable storage medium is provided, in which at least one instruction or at least one program is stored, and the at least one instruction or the at least one program is loaded and executed by a processor to implement the vehicle automatic driving decision method as described above.

[0047] On the other hand, a computer program product or a computer program is provided, the computer program product or the computer program includes computer instructions, the computer instructions are stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the method provided in various optional implementations of the above-mentioned vehicle autonomous driving decision.

[0048] The vehicle automatic driving decision-making method, device, vehicle terminal and storage medium provided by the embodiments of the present application have the following technical effects:

[0049] By acquiring the current position information, current navigation information, current driving condition information and current road environment information of the road on which the target vehicle is traveling, the driving scene of the target vehicle is determined based on the current position information and current road environment information of the target vehicle, and the driving scene includes a road driving scene and an intersection driving scene; when the driving scene of the target vehicle is the road driving scene, the current driving condition information and the current road environment information of the target vehicle are processed according to the road lateral decision strategy to obtain a first lateral decision result of the target vehicle; when the driving scene of the target vehicle is the intersection driving scene, the current navigation information and the current road environment information of the target vehicle are processed according to the intersection lateral decision strategy to obtain a second lateral decision result of the target vehicle; according to the first lateral decision result and the second lateral decision result, the driving trajectory of the target vehicle is determined, thereby reducing the complexity of the automatic driving decision system under road conditions and improving the decision efficiency and accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0051] Figure 1 This is a flow chart of a vehicle automatic driving decision-making method provided in an embodiment of the present application. Figure 1 ;

[0052] Figure 2This is a flow chart of a vehicle automatic driving decision-making method provided in an embodiment of the present application. Figure 2 ;

[0053] Figure 3 It is a structural schematic diagram of a vehicle automatic driving decision-making method provided in an embodiment of the present application;

[0054] Figure 4 It is a structural schematic diagram of a road lateral decision-making strategy for updating provided in an embodiment of the present application;

[0055] Figure 5 It is a flow chart of a lateral decision-making strategy for an intersection provided in an embodiment of the present application;

[0056] Figure 6 It is a structural schematic diagram of a vehicle automatic driving decision-making device provided in an embodiment of the present application;

[0057] Figure 7 This is a hardware structure block diagram of a vehicle automatic driving decision-making on-board terminal provided in an embodiment of the present application. DETAILED DESCRIPTION

[0058] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0059] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0060] See also Figure 1 , Figure 1This is a flow chart of a vehicle automatic driving decision method provided by an embodiment of the present application. The method obtains the current position information, current navigation information, current driving condition information and current road environment information of the target vehicle. When the distance between the current position information of the target vehicle and the target intersection is greater than or equal to a preset distance threshold, the driving scene of the target vehicle is determined to be a road driving scene, and the current driving condition information and current road environment information of the target vehicle are processed according to a road lateral decision strategy to obtain a first lateral decision result; when the distance between the current position information of the target vehicle and the target intersection is less than a preset distance threshold, the driving scene of the target vehicle is determined to be an intersection driving scene, and the current navigation information and current road environment information of the target vehicle are processed according to the intersection lateral decision strategy to obtain a second lateral decision result; based on the first lateral decision result and the second lateral decision result, the driving trajectory of the target vehicle is determined. The vehicle automatic driving decision-making method of the present invention can adopt corresponding lateral decision-making strategies according to the driving scenarios, adopt the road lateral decision-making strategy in the road driving scenario, and adopt the intersection lateral decision-making strategy in the intersection driving scenario, thereby avoiding the excessive complexity of the model in the road driving scenario caused by adopting the intersection lateral decision-making strategy in the road driving scenario, and improving the efficiency and accuracy of the vehicle automatic driving decision-making.

[0061] This specification provides method operation steps such as embodiments or flow charts, but may include more or fewer operation steps based on routine or non-creative work. The order of steps listed in the embodiments is only one way of executing the steps among many orders, and does not represent the only order of execution. When the actual system or server product is executed, it can be executed in the order of the method shown in the embodiments or the drawings or in parallel (for example, in a parallel processing or multi-threaded processing environment). Specifically, Figure 1 As shown, the method may include:

[0062] S101: Acquire current position information, current navigation information, current driving condition information of a target vehicle and current road environment information of a road on which the target vehicle is traveling;

[0063] S102: Determine a driving scene of the target vehicle based on the current position information of the target vehicle and the current road environment information, where the driving scene includes a road driving scene and an intersection driving scene;

[0064] S103: When the driving scenario of the target vehicle is the road driving scenario, the current driving condition information and the current road environment information of the target vehicle are processed according to a road lateral decision strategy to obtain a first lateral decision result of the target vehicle;

[0065] S104: When the driving scene of the target vehicle is the intersection driving scene, the current navigation information and the current road environment information of the target vehicle are processed according to the intersection lateral decision strategy to obtain a second lateral decision result of the target vehicle;

[0066] S105: Determine the driving trajectory of the target vehicle according to the first lateral decision result and / or the second lateral decision result.

[0067] In an optional embodiment of the present application, as Figure 2 As shown, including:

[0068] S201: Acquire current position information, current navigation information, current driving condition information of a target vehicle and current road environment information of a road on which the target vehicle is traveling;

[0069] Specifically, the current location information of the target vehicle can be obtained through the global positioning system (GPS); the current driving condition information of the target vehicle includes but is not limited to the speed, acceleration, heading angle, center x coordinate, center y coordinate, length and width of the target vehicle based on the rectangular frame, etc., which can be obtained through sensors installed on the car; the current navigation information includes the navigation path planning information of the target vehicle, which can be obtained through the on-board navigation system; the current road environment information includes the surrounding obstacles of the target vehicle (such as vehicles, pedestrians, buildings, etc.) and road information (such as drivable areas, lane lines, traffic signs, traffic lights, etc.), which can be obtained by installing sensors (such as cameras, lidars, millimeter-wave radars, etc.) on the target vehicle.

[0070] S202: Determine a driving scenario of the target vehicle based on the current position information and current road environment information of the target vehicle.

[0071] In the embodiments of the present application, the driving conditions of a vehicle on a road are divided into road driving conditions and intersection driving conditions. Correspondingly, the driving scenarios of the vehicle defined in the present application include road driving scenarios and intersection driving scenarios. Different lateral decision-making strategies are adopted for different driving scenarios. For example, a Markov decision method is adopted for road driving scenarios, and a finite state machine decision method is adopted for intersection driving scenarios with strong rule constraints. The advantages of the two types of decision methods are fully utilized to make decisions more efficient and intelligent.

[0072] In an embodiment of the present application, optionally, a specific method for determining the driving scene of the target vehicle is: determining whether the distance between the current position information of the target vehicle and the target intersection is greater than or equal to a preset distance threshold.

[0073] In an embodiment of the present application, optionally, the current road environment information includes target intersection information of the target vehicle, the target intersection including a front target intersection and a rear target intersection, the front target intersection being an intersection located in front of the target vehicle in the direction of travel of the target vehicle, the rear target intersection being an intersection located behind the target vehicle in the direction of travel of the target vehicle, and the preset distance threshold including a first distance threshold and a second distance threshold.

[0074] In an embodiment of the present application, the current driving scenario of the target vehicle is determined by judging the distance between the target vehicle and the target front intersection or the target rear intersection, and a corresponding lateral decision-making strategy is adopted according to the current driving scenario.

[0075] S203: When the distance between the current position information of the target vehicle and the target intersection is greater than or equal to a preset distance threshold, the driving scene of the target vehicle is determined to be a road driving scene, and a road lateral decision strategy is obtained. The current driving condition information and the current road environment information of the target vehicle are processed according to the road lateral decision strategy to obtain a first lateral decision result of the target vehicle.

[0076] In an embodiment of the present application, when the distance between the current position information of the target vehicle and the target intersection ahead is greater than or equal to the first distance threshold or the distance between the current position information of the target vehicle and the target intersection behind is greater than or equal to the second distance threshold, the target vehicle is judged to be in a road driving scene, and the target vehicle obtains a road lateral decision strategy during the automatic driving process, and processes the current driving condition information and the current road environment information of the target vehicle according to the obtained road lateral decision strategy to obtain the first lateral decision result of the target vehicle.

[0077] S204: When the distance between the current position information of the target vehicle and the target intersection exceeds the preset distance threshold, the driving scene of the target vehicle is determined to be an intersection driving scene, and the intersection lateral decision strategy is obtained. The current navigation information and current road environment information of the target vehicle are processed according to the intersection lateral decision strategy to obtain a second lateral decision result of the target vehicle.

[0078] In an embodiment of the present application, when the distance between the current position information of the target vehicle and the target intersection ahead is less than the first distance threshold or the distance between the current position information of the target vehicle and the target intersection behind is less than the second distance threshold, the target vehicle is judged to be in an intersection driving scene, and the target vehicle obtains an intersection lateral decision strategy during the automatic driving process, and processes the current navigation information and current road environment information of the target vehicle according to the obtained intersection lateral decision strategy to obtain a second lateral decision result of the target vehicle.

[0079] S205: Determine the driving trajectory of the target vehicle according to the first lateral decision result and / or the second lateral decision result.

[0080] Optionally, a method for determining the driving trajectory of the target vehicle according to the first lateral decision result and / or the second lateral decision result is:

[0081] The first lateral decision result and / or the second lateral decision result are used as the target lateral decision result of the target vehicle; based on the target lateral decision result, the opening of the throttle or brake pedal that needs to be controlled when the vehicle maintains a safe distance from surrounding obstacles during the execution of the target lateral decision result is determined, thereby determining the target longitudinal decision result; based on the target lateral decision result and the target longitudinal decision result, the target vehicle is subjected to longitudinal and lateral dynamic coupling to determine the driving trajectory of the target vehicle.

[0082] Figure 3 A flowchart of a feasible vehicle autonomous driving decision-making method for this application is shown below: Figure 3 As shown: The vehicle autonomous driving decision system includes five parts: map positioning, perception prediction, route navigation, decision planning, and vehicle control. Among them, the decision planning module includes a scene scheduling module, a lateral decision module, a comprehensive decision module and a planning module. The lateral decision module is divided into road scene lateral decision and intersection scene lateral decision.

[0083] The scene scheduling module monitors the current position information of the target vehicle in real time according to the map positioning module, sets the first distance threshold s1 and the second distance threshold s2. If the target vehicle is traveling along the road, the current position of the target vehicle is greater than the first distance threshold s1 from the target intersection in front, and the driving scene of the target vehicle is determined to be a road driving scene, then the road scene lateral decision is maintained, and the road lateral decision strategy is adopted to make a lateral decision result, otherwise the intersection scene lateral decision is entered, and the intersection lateral decision is made. The lateral decision result is made using the intersection lateral decision strategy. When the target vehicle leaves the intersection, when it is determined that the current position information of the target vehicle is less than the second distance threshold s2 from the rear target intersection, the intersection scene lateral decision is maintained, otherwise the road scene lateral decision is entered.

[0084] The lateral decision module calculates and solves the road lateral decision strategy and the intersection lateral decision strategy, and sends the lateral decision results to the comprehensive decision module. The comprehensive decision module makes a longitudinal decision based on the lateral decision results, and coordinates the lateral decision results with the longitudinal decision results and sends them to the planning module. Finally, the trajectory is generated and released to the outside to achieve automatic driving of the target vehicle.

[0085] In the embodiment of the present application, the road lateral decision strategy can be a pre-trained road lateral decision model. The target vehicle receives the road lateral decision model sent from the cloud server, uses the received road lateral decision model as the road lateral decision strategy, processes the current driving condition information and the current road environment information of the target vehicle according to the road lateral decision model, and obtains the first lateral decision result of the target vehicle.

[0086] An optional method for training a road lateral decision model is to train a preset Markov reinforcement learning model based on sample data of a sample vehicle, wherein the sample data includes sample driving condition information of the sample vehicle and sample road environment information of the road on which the sample vehicle is traveling, wherein the sample driving condition information includes the motion state, spatial state and lateral behavior state of the sample vehicle, and the sample vehicle can be a target vehicle or other autonomous driving vehicles.

[0087] The Markov-based road lateral decision model can be expressed as a four-tuple (S, A, P, R): where S is the state space, S = [S adv , S1, S2, ..., S n ] T , driving condition information S of the autonomous driving vehicle adv Including the motion state and spatial state of the autonomous driving vehicle itself, S adv =[v,a,x,y,θ,l,w] T , v, a, θ are the speed, acceleration, and heading angle of the autonomous driving vehicle, respectively, representing the vehicle's motion state; x, y, l, and w are the center x coordinate, center y coordinate, length, and width of the rectangular box of the autonomous driving vehicle, respectively, representing the vehicle's spatial state. The state of the other i-th obstacle vehicle S i =[t, b, v, a, x, y, θ, l, w] T , where t and b are the vehicle type and predicted driving behavior respectively, S i The road environment information of the road on which the autonomous driving vehicle is traveling can be represented.

[0088] A is the action space, which represents the first lateral decision result of the autonomous driving vehicle. The lateral driving action instructions for road driving are defined as a discrete set, A = [LK, LC, RC, LN, RN], where LK, LC, RC, LN, and RN respectively represent the first lateral decision results of the vehicle, which are lane keeping, lane change to the left, lane change to the right, avoid to the left, and avoid to the right.

[0089] P(s′|s) is a state transition probability model, which describes the evolution of the driving state of the autonomous driving vehicle over time. It indicates the probability that the autonomous driving vehicle changes from the current lateral state to the next lateral state after executing the corresponding road lateral driving action command:

[0090]

[0091] Among them, P(s′|s) represents the probability that the autonomous driving vehicle transfers from the current state s to the state s' at the next moment;

[0092] P adv (s′ a |s a , a) means that if an action a is executed, the autonomous driving vehicle changes from the current state s a Transfer to the next state s a 'Probability;

[0093] P(s′ i |s i ) indicates that other vehicles are in the current state s i Transfer to the next state s i The probability that the autonomous vehicle is in the state s at a certain action A a It can be expressed as:

[0094] v a =v+a a τ

[0095] a a =a+jτ

[0096] x a =x+(v+aτ)cos(θ+Δθ)

[0097] y a =y+(v+aτ)cos(θ+Δθ)

[0098] θ a =θ+Δθ

[0099] l a = l

[0100] w a =w

[0101] Where τ is the decision cycle, j is the average jerk in the past several decision cycles, and Δθ is the heading angle increment. i , since the perception prediction cycle is generally not less than the decision planning calculation cycle, it can be considered that the predicted driving behavior and vehicle prediction type in the driving action generation cycle are unchanged, so the state transition probability model P of other vehicles i Please refer to the above-mentioned autonomous driving vehicle.

[0102] R is a weighted reward function, which is an evaluation criterion for the vehicle's road lateral decision-making strategy. It mainly considers the five dimensions of safety, traffic rules, comfort, traffic efficiency and completion of the autonomous driving vehicle during the autonomous driving process:

[0103] R=γ1f s +γ2f r +γ3f c +γ4f e +γ5f t

[0104] where γ i is the weight coefficient of the reward function in different dimensions, and 0≤γ i ≤1; f s is the safety factor reward function, which mainly considers the distance between the autonomous driving vehicle and other vehicles during driving. The greater the distance, the greater the reward function value; r is the reward function for the autonomous driving vehicle to comply with traffic rules. For driving behaviors that violate traffic rules during driving, a fixed penalty value is given and the reward function value decreases; f c The impact of the lateral acceleration and longitudinal acceleration of the autonomous driving vehicle is mainly considered. The greater the impact, the lower the reward function value; f e The speed of the autonomous vehicle is mainly considered. The closer the speed of the autonomous vehicle is to the cruising speed, the greater the reward function value; f t The reward function for completing the autonomous driving task mainly considers whether the autonomous driving vehicle reaches the effective range of the target point during the autonomous driving process. If it enters the effective range, a fixed reward value is given to increase the reward function value. Otherwise, a penalty value is given to reduce the reward function value.

[0105] After the driving condition information and road environment information of the autonomous driving vehicle are input into the road lateral decision model, after the model autonomous driving vehicle executes different road lateral driving action instructions, the autonomous driving vehicle has a certain probability of changing from the current lateral state to the next lateral state, and also has a certain probability of maintaining the current lateral state. Each lateral state will correspond to a reward value, which can be calculated by a reward function; optionally, the road lateral driving action instruction corresponding to a reward value greater than a preset reward threshold can be used as the first lateral decision result output by the model.

[0106] The road lateral decision model in the embodiment of the present application uses the safety, traffic rules, comfort and traffic efficiency of the autonomous driving vehicle during the autonomous driving process as the evaluation criteria for the road lateral decision strategy. The evaluation effect of the road lateral decision model is more comprehensive, which improves the driving safety and driving comfort of the autonomous driving vehicle and improves the riding experience.

[0107] In order to ensure that the road lateral decision strategy has better applicability, the target vehicle can also receive the updated road lateral decision model sent by the cloud server during the autonomous driving process, and use the updated road lateral decision model as the road lateral decision strategy, such as Figure 4 As shown:

[0108] The on-board sensor obtains the current driving condition information and current road environment information of the target vehicle in real time, and continuously sends the current driving condition information and current road environment information of the target vehicle to the road-side communication device through the on-board computing unit and the on-board communication unit, and transmits it to the cloud server through the road-side communication device. The learning system on the cloud server updates the sample data of the sample vehicle based on the current driving condition information and current road environment information of the target vehicle collected in real time, retrains and learns the preset Markov model according to the updated data, iteratively optimizes, and feeds back the updated road lateral decision model to the autonomous driving vehicle computing unit for standby. When the vehicle's autonomous driving system is started next time, the updated road lateral decision model will be used as the road lateral decision strategy.

[0109] In the embodiment of the present application, the lateral decision strategy at the intersection is a lateral decision strategy based on a finite state machine. A feasible method for constructing the lateral decision strategy at the intersection is:

[0110] Obtain preset navigation information of a preset vehicle, preset road environment information of a road on which the preset vehicle is traveling, a preset initial lateral decision result, and a preset target lateral decision result; establish a mapping relationship between the preset navigation information, the preset road environment information, the preset initial lateral decision result, and the preset target lateral decision result; and use the mapping relationship as a lateral decision strategy for an intersection.

[0111] Figure 5 is the structural diagram of the lateral decision strategy at the intersection, such as Figure 5 As shown: The lateral decision results of the intersection lateral decision strategy include the preparation state S PR , go straight S GS , turn left S TL , turn right S TR , U-turn S UT , where there is only a transfer relationship between the initial lateral decision result and the target lateral decision result. When the corresponding conditions are met, the lateral state of the target vehicle is transferred from the initial lateral decision result to the target lateral decision result. The transfer relationship between the initial lateral decision result and the target lateral decision result is shown in Table 1 below:

[0112] Table 1 - Transfer relationship table of initial horizontal decision results and target horizontal decision results

[0113]

[0114]

[0115] In this application, when the target vehicle is in an intersection condition, the intersection lateral decision strategy is used according to step S204 to process the current navigation information and the current road environment information of the target vehicle to obtain a second lateral decision result of the target vehicle:

[0116] Obtain the current initial lateral decision result of the target vehicle; obtain the preset lateral decision result corresponding to the current navigation information, the current road environment information and the current initial lateral decision result in Table 1 according to the pre-constructed mapping relationship; and use the preset lateral decision result corresponding to the current navigation information, the current road environment information and the current initial lateral decision result as the second lateral decision result.

[0117] The present application provides a vehicle automatic driving decision method, which obtains the current position information, current navigation information, current driving condition information and current road environment information of the road on which the target vehicle is traveling. When the distance between the current position information of the target vehicle and the target intersection is greater than or equal to a preset distance threshold, the driving scene of the target vehicle is determined to be a road driving scene, and the current driving condition information and current road environment information of the target vehicle are processed according to the road lateral decision strategy to obtain a first lateral decision result; when the distance between the current position information of the target vehicle and the target intersection is less than the preset distance threshold, the driving scene of the target vehicle is determined to be an intersection driving scene, and the current navigation information and current road environment information of the target vehicle are processed according to the intersection lateral decision strategy to obtain a second lateral decision result; and the driving trajectory of the target vehicle is determined according to the first lateral decision result and the second lateral decision result. Different lateral decision methods are used for different driving scenes, which reduces the complexity of the automatic driving decision system under road conditions and improves the efficiency and accuracy of automatic driving decisions.

[0118] On the other hand, the embodiment of the present application also provides a vehicle automatic driving decision-making device, Figure 6 is a schematic diagram of the structure of a vehicle automatic driving decision-making device provided in an embodiment of the present application, such as Figure 6 As shown, the device comprises:

[0119] The acquisition module 601 is used to acquire the current position information, current navigation information, current driving condition information of the target vehicle and the current road environment information of the road on which the target vehicle is traveling;

[0120] A scene determination module 602 is used to determine the driving scene of the target vehicle based on the current position information and the current road environment information of the target vehicle, wherein the driving scene includes a road driving scene and an intersection driving scene;

[0121] The first lateral decision module 603 is used to process the current driving condition information and the current road environment information of the target vehicle according to the road lateral decision strategy when the driving scene of the target vehicle is a road driving scene, so as to obtain a first lateral decision result of the target vehicle;

[0122] The second lateral decision module 604 is used to process the current navigation information and the current road environment information of the target vehicle according to the intersection lateral decision strategy when the driving scene of the target vehicle is an intersection driving scene, so as to obtain a second lateral decision result of the target vehicle;

[0123] The trajectory determination module 605 is used to determine the driving trajectory of the target vehicle according to the first lateral decision result and / or the second lateral decision result.

[0124] The device and method embodiments in the embodiments of the present application are based on the same application concept.

[0125] The vehicle automatic driving decision method and device provided by the above-mentioned application are applied to the field of vehicle automatic driving technology. By acquiring the current position information, current navigation information, current driving condition information and current road environment information of the target vehicle in real time, it is judged whether the vehicle is in a road condition or an intersection condition. A road lateral decision strategy is adopted in road conditions, and an intersection lateral decision strategy is adopted in intersection conditions, thereby reducing the complexity of the decision system and improving the decision efficiency and accuracy.

[0126] An embodiment of the present invention provides a vehicle-mounted terminal, which includes a processor and a memory, in which at least one instruction or at least one program is stored, and the at least one instruction or the at least one program is loaded and executed by the processor to implement the vehicle automatic driving decision method provided in the above method embodiment.

[0127] The memory can be used to store software programs and modules, and the processor executes various functional applications by running the software programs and modules stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, application programs required for functions, etc.; the data storage area can store data created according to the use of the device, etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage devices. Accordingly, the memory can also include a memory controller to provide the processor with access to the memory.

[0128] The method embodiment provided in the embodiment of the present invention can be executed on a vehicle-mounted terminal. Figure 7As shown, the internal structure of the vehicle-mounted terminal may include but is not limited to: a processor, a network interface and a memory, wherein the processor, the network interface and the memory may be connected via a bus or other means.

[0129] Among them, the processor (or CPU (Central Processing Unit)) is the computing core and control core of the computer device. The network interface may optionally include a standard wired interface, a wireless interface (such as WI-FI, a mobile communication interface, etc.). Memory is a memory device in a computer device for storing programs and data. It is understandable that the memory here can be a high-speed RAM storage device or a non-volatile memory device (non-volatile memory), such as at least one disk storage device; optionally, it can also be at least one storage device located away from the aforementioned processor. The memory provides a storage space, which stores the operating system of the electronic device, including but not limited to: Windows system (an operating system), Linux (an operating system), Android (Android, a mobile operating system) system, IOS (a mobile operating system) system, etc., and the present invention is not limited to this; and, in the storage space, one or more instructions suitable for being loaded and executed by the processor are also stored, and these instructions can be one or more computer programs (including program codes). In the embodiment of this specification, the processor loads and executes one or more instructions stored in the memory to implement the vehicle automatic driving decision method provided in the above method embodiment.

[0130] An embodiment of the present invention also provides a computer-readable storage medium, which can be set in a terminal to store at least one instruction, at least one program, code set or instruction set related to implementing a vehicle automatic driving decision-making method. The at least one instruction, the at least one program, the code set or instruction set is loaded and executed by the processor to implement the vehicle automatic driving decision-making method provided by the above-mentioned method embodiment.

[0131] Optionally, in this embodiment, the above-mentioned storage medium may include but is not limited to: a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and other media that can store program codes.

[0132] It should be noted that the above-mentioned sequence of the embodiments of the present application is for description only and does not represent the advantages and disadvantages of the embodiments. The above-mentioned specific embodiments of this specification are described. Other embodiments are within the scope of the attached claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0133] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0134] A person skilled in the art will understand that all or part of the steps to implement the above embodiments may be accomplished by hardware or by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a disk or an optical disk, etc.

[0135] The above description is only a preferred embodiment of the present invention and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present application.

Claims

1. A vehicle automatic driving decision-making method, characterized in that: include: Acquiring current location information, current navigation information, current driving condition information of a target vehicle and current road environment information of a road on which the target vehicle is traveling; Based on the current position information of the target vehicle and the current road environment information, determining the driving scene of the target vehicle, the driving scene including the road driving scene and the intersection driving scene; When the driving scene of the target vehicle is the road driving scene, the current driving condition information and the current road environment information of the target vehicle are processed according to a road lateral decision strategy to obtain a first lateral decision result of the target vehicle; the road lateral decision strategy is a pre-trained road lateral decision model, and the pre-trained road lateral decision model is established based on a state space, an action space, a state probability transfer function and a reward function; When the driving scene of the target vehicle is the intersection driving scene, the current navigation information and the current road environment information of the target vehicle are processed according to the intersection lateral decision strategy to obtain a second lateral decision result of the target vehicle; the intersection lateral decision model is a lateral decision strategy based on a finite state machine, and the lateral decision strategy based on a finite state machine is obtained by establishing a mapping relationship between preset navigation information, preset road environment information and preset initial lateral decision result and a preset target lateral decision result; The driving trajectory of the target vehicle is determined according to the first lateral decision result and / or the second lateral decision result.

2. The vehicle automatic driving decision-making method according to claim 1, characterized in that: The determining the driving scene of the target vehicle based on the current position information of the target vehicle and the current road environment information includes: The current road environment information includes information of a front target intersection and a rear target intersection of the target vehicle; When the distance between the current position information of the target vehicle and the front target intersection is greater than or equal to a first distance threshold or the distance between the current position information of the target vehicle and the rear target intersection is greater than or equal to a second distance threshold, the driving scene of the target vehicle is determined to be the road driving scene.

3. The vehicle automatic driving decision method according to any one of claims 1-2, characterized in that: The determining of the driving scene of the target vehicle based on the current position information and the current road environment information of the target vehicle includes: The current road environment information includes information about the front target intersection and the rear target intersection of the target vehicle; When the distance between the current position information of the target vehicle and the front target intersection is less than a first distance threshold or the distance between the target position information and the rear target intersection is less than a second distance threshold, the driving scene of the target vehicle is determined to be the intersection driving scene.

4. The vehicle automatic driving decision method according to claim 3, characterized in that: The processing of the current driving condition information and the current road environment information of the target vehicle according to the road lateral decision strategy to obtain a first lateral decision result of the target vehicle includes: Receiving the road lateral decision model sent by the server; Using the received road lateral decision model as the road lateral decision strategy; Processing the current driving condition information and the current road environment information of the target vehicle according to the road lateral decision model to obtain the first lateral decision result of the target vehicle; Among them, the road lateral decision model is obtained by training a preset model based on sample data of a sample vehicle. The preset model is a model based on Markov reinforcement learning. The sample data includes sample driving condition information of the sample vehicle and sample road environment information of the road on which the sample vehicle is traveling.

5. The vehicle automatic driving decision-making method according to claim 4, characterized in that: The method further comprises: Receiving the updated road lateral decision model sent by the server; Using the updated road lateral decision model as the road lateral decision strategy; Among them, the updated road lateral decision model is obtained by training the preset model based on the updated sample data; the updated sample data is obtained by updating the sample data based on the current driving condition information of the target vehicle and the current road environment information.

6. The vehicle automatic driving decision-making method according to claim 1, characterized in that: The method further comprises: Acquiring preset navigation information of the preset vehicle, preset road environment information of the road on which the preset vehicle is traveling, the preset initial lateral decision result, and the preset target lateral decision result; Establishing a mapping relationship between the preset navigation information, the preset road environment information and the preset initial lateral decision result, and the preset target lateral decision result; The mapping relationship is used as the lateral decision-making strategy for the intersection.

7. The vehicle automatic driving decision-making method according to claim 6, characterized in that: The step of processing the current navigation information and the current road environment information of the target vehicle according to the intersection lateral decision strategy to obtain a second lateral decision result of the target vehicle includes: Obtaining a current initial lateral decision result of the target vehicle; According to the mapping relationship, obtaining a preset lateral decision result corresponding to the current navigation information, the current road environment information and the current initial lateral decision result; The preset lateral decision result corresponding to the current navigation information, the current road environment information and the current initial lateral decision result is used as the second lateral decision result.

8. The vehicle automatic driving decision-making method according to claim 1, characterized in that: Determining the driving trajectory of the target vehicle according to the first lateral decision result and / or the second lateral decision result includes: Using the first lateral decision result and / or the second lateral decision result as a target lateral decision result of the target vehicle; Determine the target vertical decision result based on the target horizontal decision result; The driving track of the target vehicle is determined according to the target lateral decision result and the target longitudinal decision result.

9. A vehicle automatic driving decision-making device, characterized in that: include: An acquisition module, used to acquire the current position information, current navigation information, current driving condition information of the target vehicle and the current road environment information of the road on which the target vehicle is traveling; A scene determination module, used to determine the driving scene of the target vehicle based on the current position information of the target vehicle and the current road environment information, wherein the driving scene includes a road driving scene and an intersection driving scene; A first lateral decision module is used to process the current driving condition information and the current road environment information of the target vehicle according to a road lateral decision strategy when the driving scenario of the target vehicle is the road driving scenario, so as to obtain a first lateral decision result of the target vehicle; the road lateral decision strategy is a pre-trained road lateral decision model, and the pre-trained road lateral decision model is established based on a state space, an action space, a state probability transfer function and a reward function; A second lateral decision module is used for processing the current navigation information and the current road environment information of the target vehicle according to the lateral decision strategy at the intersection when the driving scene of the target vehicle is the driving scene at the intersection, so as to obtain a second lateral decision result of the target vehicle; the lateral decision model at the intersection is a lateral decision strategy based on a finite state machine, and the lateral decision strategy based on a finite state machine is obtained by establishing a mapping relationship between preset navigation information, preset road environment information and preset initial lateral decision result and a preset target lateral decision result; The trajectory determination module is used to determine the driving trajectory of the target vehicle according to the first lateral decision result and / or the second lateral decision result.

10. A vehicle-mounted terminal, characterized in that: The vehicle-mounted terminal includes a processor and a memory, wherein the memory stores at least one instruction or at least one program, and the at least one instruction or the at least one program is loaded and executed by the processor to implement the vehicle automatic driving decision method as described in any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that: The storage medium stores at least one instruction or at least one program, and the at least one instruction or the at least one program is loaded and executed by the processor to implement the vehicle automatic driving decision method as described in any one of claims 1 to 8.

Citation Information

Patent Citations

  • Driving behavior decision system of automatic driving public traffic vehicle

    CN107272687A