Surrounding vehicle trajectory prediction method driven by behavior cognition in multi-vehicle interference scenarios

Abstract

This invention relates to the field of intelligent driving technology and discloses a method for predicting the trajectory of surrounding vehicles driven by behavior cognition in multi-vehicle interference scenarios. The method includes: identifying a target vehicle with primary interference factors and its first and second behavioral characteristics among the surrounding vehicles based on acquired urban road driving data; determining the target's emotional characteristics based on environmental information and driver status monitoring information of the target vehicle; obtaining the target vehicle's behavioral intention based on the first and second behavioral characteristics, the target's emotional characteristics corresponding to the target vehicle's driving behavior, and road geometric features; determining the target road condition information based on acquired road geometric features, road dynamic elements, multi-vehicle interaction interference motion characteristics, and multi-vehicle interference interaction intention information; and obtaining the trajectory prediction result based on the behavioral intention and the target road condition information. This invention achieves accurate prediction of the operating trajectories of surrounding vehicles.

Description

Technical Field

[0001] The present invention relates to the field of intelligent driving technology, and in particular to a method for predicting surrounding vehicle trajectories driven by behavior cognition in a multi-vehicle interference scenario. Background Art

[0002] The vehicle driving environment on urban roads is complex, and the behavior of traffic participants is highly uncertain. They often encounter scenes of interference from multiple vehicles around them. For example, when the traffic flow density is high, due to the different driving behaviors and driving speeds of surrounding vehicles (number ≥ 2, close to the safe vehicle distance range), the interaction between vehicles affects the driving of the vehicle, especially when there are multiple vehicles on the road section. For example, there are parallel vehicles in the left and right adjacent lanes, and the vehicle in front jointly affects the operation of the vehicle. The interference of multiple vehicles at intersections is more complicated. For example, the vehicle going straight may encounter a right-turning vehicle, and there may be parallel vehicles on the left or right side of the vehicle. It can be seen that on urban roads, there will be many complex situations that affect the driving of the vehicle in the case of multiple vehicle interference. Therefore, it is more important to predict the trajectory of surrounding vehicles.

[0003] In the related art, the trajectory prediction of surrounding vehicles is to determine the main lane and adjacent lanes of the main vehicle based on the lane lines; determine the target vehicles around the main vehicle from the main lane and adjacent lanes based on the preset environmental model; obtain the path prediction type, predicted speed and lateral path characteristics of the target vehicle; analyze the motivation and intention type of the target vehicle based on the path prediction type, predicted speed and lateral path characteristics; and predict the driving trajectory of the target vehicle based on the motivation and intention type. However, in the related art, the information of the target vehicle is determined only based on the environmental model, so that the information of the target vehicle obtained is not accurate, resulting in the motivation and intention type obtained based on the information of the target vehicle being inaccurate, making it difficult to fully describe the trajectory uncertainty caused by the interference interaction of multiple vehicles, and failing to fit the running trajectory of the surrounding vehicles in actual situations. Summary of the invention

[0004] In view of this, the present invention provides a behavior cognition-driven surrounding vehicle trajectory prediction method in a multi-vehicle interference scenario to solve the problem of inaccurate prediction of the running trajectory of surrounding vehicles.

[0005] In a first aspect, the present invention provides a method for predicting surrounding vehicle trajectories driven by behavioral cognition in a multi-vehicle interference scenario, comprising: determining a target vehicle with a main interference factor among the surrounding vehicles of the own vehicle based on acquired urban road driving data; wherein the main interference factor represents the interference behavior and the interfering vehicle that have a direct impact on the safe driving of the own vehicle in the multi-vehicle interference scenario; determining a first behavioral characteristic and a second behavioral characteristic based on the urban road driving data; wherein the first behavioral characteristic represents at least one of the acceleration change, speed change, direction angle, yaw angular velocity and driving intention change generated by the own vehicle under the interactive interference behavior of the surrounding vehicles; and the second behavioral characteristic represents at least one of the acceleration change, speed change, direction angle, yaw angular velocity and driving intention change generated by the target vehicle during driving; according to The target emotion characteristics corresponding to the driving behavior of the target vehicle are determined based on the environmental information and the driver status monitoring information of the target vehicle; the first behavior characteristics, the second behavior characteristics, the target emotion characteristics corresponding to the driving behavior of the target vehicle and the acquired road geometry characteristics are input into the trained surrounding vehicle behavior intention model to obtain the behavior intention of the target vehicle; wherein the road geometry characteristics represent the static characteristics of the road; according to the acquired road geometry characteristics, road dynamic elements, multi-vehicle interaction interference motion characteristics and multi-vehicle interference interaction intention information, an environmental dynamic model is constructed; according to the environmental dynamic model, the target road condition information of the target vehicle is determined; the behavior intention and the target road condition information are input into the trained trajectory prediction model to obtain the trajectory prediction result of the target vehicle, wherein the target road condition information represents the dynamic condition information of the road.

[0006] The present invention determines the target vehicle among the surrounding vehicles of the self-vehicle that has a direct impact on the safe driving of the self-vehicle in the scenario of multi-vehicle interference according to the acquired urban road driving data, determines the first behavioral characteristics of the self-vehicle and the second behavioral characteristics of the target vehicle according to the urban road driving data, determines the target emotional characteristics corresponding to the driving behavior of the target vehicle according to the environmental information and the driver status monitoring information of the target vehicle, and obtains the behavioral intention of the target vehicle according to the first behavioral characteristics, the second behavioral characteristics, the target emotional characteristics corresponding to the driving behavior of the target vehicle and the acquired road geometric characteristics. The present invention considers the influence of the driver's emotional characteristics on the driving intention, establishes the representation relationship between the driver's emotions and the movement changes of the surrounding vehicles, and more accurately infers the driving intention of the surrounding vehicles. The present invention takes the first behavioral characteristics of the self-vehicle, the second behavioral characteristics of the target vehicle, the target emotional characteristics corresponding to the driving behavior of the target vehicle, and the road geometric characteristics as the influencing factors of the behavioral intention of the target vehicle, and the consideration range is wide, so that the behavioral intention obtained is more accurate. In the interactive process facing the multi-vehicle interference of urban roads, the present invention is particularly aimed at the interactive behavior law of the surrounding vehicles of the multi-vehicle interference, and accurately analyzes and determines the surrounding vehicle motion characteristics and driving intention of the main interference factor. The present invention constructs an environmental dynamic model based on the acquired road geometry features, road dynamic elements, multi-vehicle interactive interference motion characteristics, and multi-vehicle interference interactive intention information; according to the environmental dynamic model, the target road condition information of the target vehicle is determined, and more factors are considered when constructing the dynamic environmental model, so that the target road condition information is more accurate and more in line with the actual road condition information. The present invention inputs the behavioral intention and the target road condition information into the trained trajectory prediction model to obtain the trajectory prediction result of the target vehicle. Since the behavioral intention and the target road condition information are more accurate, the trajectory prediction result obtained according to the behavioral intention and the target road condition information is more accurate. The present invention considers the motion characteristics of the self-vehicle and the target vehicle and the emotional characteristics of the driver of the target vehicle, and significantly improves the trajectory prediction accuracy of the target vehicle of the main interference factor in the multi-vehicle interference scene.

[0007] In an optional embodiment, based on the acquired urban road driving data, a target vehicle with a main interference factor among the surrounding vehicles is determined, including: analyzing the multi-vehicle interference behavior of the surrounding vehicles based on the urban road driving data; wherein the multi-vehicle interference behavior represents the type of impact on the driving of the own vehicle when the driving behavior and driving speed of the surrounding vehicles change; based on the multi-vehicle interference behavior, the target vehicle with the main interference factor among the surrounding vehicles is determined.

[0008] The present invention analyzes the multi-vehicle interference behavior that affects the driving of the own vehicle when the driving behavior and driving speed of the surrounding vehicles change based on the urban road driving data, and determines the target vehicle with the main interference factor among the surrounding vehicles according to the multi-vehicle interference behavior, and selects the vehicle that directly affects the driving of the own vehicle from the surrounding vehicles as the target vehicle. The target vehicle is mainly studied in the subsequent research, so that the trajectory prediction result of the target vehicle is more representative.

[0009] In an optional implementation, determining the first behavior characteristic and the second behavior characteristic based on urban road driving data includes: determining the first behavior characteristic of the ego vehicle and the second behavior characteristic of the target vehicle based on multi-vehicle interference behavior.

[0010] The present invention determines a first behavior characteristic of an ego vehicle and a second behavior characteristic of a target vehicle according to the multi-vehicle interference behavior, so as to facilitate subsequent prediction of the intended behavior according to the first behavior characteristic of the ego vehicle and the second behavior characteristic of the target vehicle.

[0011] In an optional implementation, the multi-vehicle interference behavior includes interference from surrounding vehicles when driving on a road section, interference from surrounding vehicles when driving at an intersection, and interference from surrounding vehicles when driving on a ramp.

[0012] In an optional embodiment, a target emotional feature corresponding to the target vehicle's driving behavior is determined based on environmental information and driver status monitoring information, including: selecting a target emotional feature corresponding to the emotional state corresponding to the target vehicle's driving behavior from preset emotional features based on environmental information and driver status monitoring information; wherein the preset emotional features include multiple emotional states.

[0013] The present invention determines the target emotional features corresponding to the emotional state corresponding to the driving behavior of the target vehicle, and uses the target emotional features as factors for determining the behavioral intention of the target vehicle, which is more in line with the actual situation and the obtained behavioral intention is more accurate.

[0014] In an optional implementation, the trajectory prediction result includes a speed prediction result and a turning angle prediction result of the target vehicle.

[0015] The trajectory prediction result of the present invention includes the speed prediction result and the turning angle prediction result of the target vehicle, and the speed prediction result and the turning angle prediction result are used to predict the longitudinal and lateral trajectory of the target vehicle, making the trajectory prediction result more accurate.

[0016] In an optional embodiment, the behavior cognition-driven surrounding vehicle trajectory prediction method in a multi-vehicle interference scenario also includes a process of training a surrounding vehicle behavior intention model, and the process of training the surrounding vehicle behavior intention model includes: establishing a surrounding vehicle behavior intention model based on a first preset algorithm; inputting the acquired urban road driving data into the surrounding vehicle behavior intention model, and training the surrounding vehicle behavior intention model; wherein the input of the surrounding vehicle behavior intention model is the urban road driving data, and the output of the surrounding vehicle behavior intention model is the behavior intention information of the surrounding vehicles.

[0017] The present invention trains a surrounding vehicle behavior intention model according to urban road driving data, and continuously improves the accuracy of the surrounding vehicle behavior intention model, so that in the application of the surrounding vehicle behavior intention model, the predicted surrounding vehicle behavior intention is more accurate.

[0018] In an optional embodiment, the behavior cognition-driven surrounding vehicle trajectory prediction method in a multi-vehicle interference scenario also includes: a process of training a trajectory prediction model, the process of training the trajectory prediction model includes: constructing a trajectory prediction model based on a second preset algorithm; inputting the acquired preset behavior intention information and preset road condition information of the surrounding vehicles into the trajectory prediction model, and training the trajectory prediction model; wherein the input of the trajectory prediction model is the preset behavior intention information and the preset road condition information, and the output of the trajectory prediction model is the trajectory prediction information of the surrounding vehicles.

[0019] The present invention trains the trajectory prediction model according to preset behavior intention information and preset road condition information, and continuously improves the accuracy of the trajectory prediction model, so that in the application of the trajectory prediction model, the predicted trajectory information is more accurate.

[0020] In a second aspect, the present invention provides a trajectory prediction model device, including: a target vehicle determination module, which is used to determine a target vehicle with a main interference factor among the surrounding vehicles of the own vehicle based on the acquired urban road driving data; wherein the main interference factor represents the interference vehicle and the interference behavior and the interference vehicle that have a direct impact on the safe driving of the own vehicle in the scenario of multi-vehicle interference; a behavior characteristic determination module, which is used to determine a first behavior characteristic and a second behavior characteristic based on the urban road driving data; wherein the first behavior characteristic represents at least one of the acceleration change, speed change, direction angle, yaw angular velocity and driving intention change generated by the own vehicle under the interactive interference behavior of the surrounding vehicles; the second behavior characteristic represents at least one of the acceleration change, speed change, direction angle, yaw angular velocity and driving intention change generated by the target vehicle during driving; an emotion feature determination module, which is used to determine the target vehicle with a main interference factor among the surrounding vehicles of the own vehicle based on the environmental information and the emotion feature determination module. The vehicle driver status monitoring information is used to determine the target emotional characteristics corresponding to the driving behavior of the target vehicle; the behavior intention determination module is used to input the first behavior characteristic, the second behavior characteristic, the target emotional characteristics corresponding to the driving behavior of the target vehicle and the acquired road geometry characteristics into the trained surrounding vehicle behavior intention model to obtain the behavior intention of the target vehicle; wherein the road geometry characteristics represent the static characteristics of the road; the target road condition information determination module is used to construct an environmental dynamic model based on the acquired road geometry characteristics, road dynamic elements, multi-vehicle interactive interference motion characteristics and multi-vehicle interference interactive intention information; according to the environmental dynamic model, the target road condition information of the target vehicle is determined; the trajectory prediction result determination module is used to input the behavior intention and the target road condition information into the trained trajectory prediction model to obtain the trajectory prediction result of the target vehicle, wherein the target road condition information represents the dynamic condition information of the road. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0022] Figure 1 is a flow chart of a method for predicting surrounding vehicle trajectories driven by behavior cognition in a multi-vehicle interference scenario according to an embodiment of the present invention;

[0023] Figure 2 is a schematic diagram of a multi-vehicle interference scenario on an urban road according to an embodiment of the present invention;

[0024] Figure 3is a flow chart of another method for predicting surrounding vehicle trajectories driven by behavior cognition in a multi-vehicle interference scenario according to an embodiment of the present invention;

[0025] Figure 4 is a flow chart of another method for predicting surrounding vehicle trajectories driven by behavior cognition in a multi-vehicle interference scenario according to an embodiment of the present invention;

[0026] Figure 5 4 is a structural block diagram of a surrounding vehicle trajectory prediction device driven by behavior cognition in a multi-vehicle interference scenario according to an embodiment of the present invention. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0028] The embodiment of the present invention provides a method for predicting surrounding vehicle trajectories driven by behavior cognition in a multi-vehicle interference scenario, which achieves the effect of accurately predicting the trajectory of a target vehicle by determining the behavior intention and target road condition information of the target vehicle.

[0029] According to an embodiment of the present invention, an embodiment of a method for predicting surrounding vehicle trajectories driven by behavior cognition in a multi-vehicle interference scenario is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in an order different from that shown here.

[0030] In this embodiment, a method for predicting surrounding vehicle trajectories driven by behavior cognition in a multi-vehicle interference scenario is provided, which can be used in a computer device. Figure 1 is a flow chart of a method for predicting surrounding vehicle trajectories driven by behavior cognition in a multi-vehicle interference scenario according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps:

[0031] Step S101, based on the acquired urban road driving data, determine the target vehicle with main interference factors among the surrounding vehicles of the own vehicle; wherein the main interference factor refers to the interference behavior and interference vehicle that have a direct impact on the safe driving of the own vehicle in the multi-vehicle interference scenario.

[0032] Among them, urban road driving data refers to real natural driving data on urban roads; urban road types include road sections, intersections, and ramps, etc.

[0033] In some optional implementations, the surrounding vehicles of the vehicle are determined based on the urban road driving data, and the vehicles that have a direct impact on the safe driving of the vehicle in a multi-vehicle interference scenario are selected from the surrounding vehicles as target vehicles.

[0034] For example, Figure 2 The figure shows a schematic diagram of an urban road environment under multi-vehicle interference. For the road section interference area, the other vehicles in the figure are the surrounding vehicles of the own vehicle, and the positions of the other vehicles directly interfere with the lane changing behavior of the own vehicle, so the target vehicles are other vehicles a, other vehicles b and other vehicles c; for the intersection interference area, the right turn behavior of other vehicles directly interferes with the straight driving of the own vehicle, so the target vehicles are other vehicles d, other vehicles e and other vehicles f.

[0035] Step S102, determining a first behavior characteristic and a second behavior characteristic based on the urban road driving data; wherein the first behavior characteristic represents at least one of the acceleration change, speed change, direction angle, yaw angular velocity and driving intention change generated by the vehicle under the interactive interference behavior of surrounding vehicles; and the second behavior characteristic represents at least one of the acceleration change, speed change, direction angle, yaw angular velocity and driving intention change generated by the target vehicle during driving.

[0036] Among them, urban road driving data includes information such as acceleration changes, speed changes, direction angles, yaw angular velocity, and driving intention changes of the vehicle and surrounding vehicles.

[0037] Wherein, illustratively, the formula of the first behavior characteristic is:

[0038] C s =[v s ,a s ,w s ,ε s ,δ s ];

[0039] Among them, C s is the first behavioral characteristic, v s is the change of vehicle speed, a s is the acceleration change of the vehicle, w s is the direction angle of the vehicle, ε s is the yaw rate of the vehicle, δ s Changes in driving intentions, including changing lanes, following, overtaking, giving way, etc.

[0040] C={C1,C2,…,C i ,…,C n};

[0041] C i =[v i,a i ,w i ,ε i ,δ i ]; i∈{1,2,…,N};

[0042] Among them, C represents the target vehicle set in the multi-vehicle interference scenario, C n represents the nth target vehicle, C i Indicates the second behavior characteristic, v i represents the speed change of the i-th vehicle, a i represents the acceleration change of the i-th vehicle, w i represents the direction angle of the i-th vehicle, ε i represents the yaw rate of the i-th vehicle, δ i Indicates the change in driving intention of the i-th vehicle, including lane changing, following, overtaking, giving way, etc.

[0043] Step S103, determining the target emotion characteristics corresponding to the driving behavior of the target vehicle according to the environmental information and the driver status monitoring information of the target vehicle.

[0044] The environmental information includes the road traffic environment and the surrounding road conditions when the vehicle is driving. The surrounding road conditions when the vehicle is driving include surrounding vehicles, and the road traffic environment includes traffic light information.

[0045] In some optional implementations, the driver status monitoring information of the target vehicle is obtained according to the driver status monitoring module of the smart cockpit.

[0046] In some optional embodiments, the behavior cognition-driven surrounding vehicle trajectory prediction method in a multi-vehicle interference scenario also includes defining the emotional characteristics of the driver of the target vehicle, and dividing the emotional characteristics of the driver of the target vehicle into good emotion, abnormal emotion, and extreme emotion.

[0047] The embodiment of the present invention selects a target emotional feature corresponding to a driving behavior corresponding to the target vehicle from the driver emotional features according to the environmental information and the driver state monitoring information of the target vehicle.

[0048] Step S104, input the first behavior characteristic, the second behavior characteristic, the target emotion characteristic corresponding to the target vehicle driving behavior, and the acquired road geometry characteristic into the trained surrounding vehicle behavior intention model to obtain the behavior intention of the target vehicle; wherein the road geometry characteristic represents the static characteristic of the road.

[0049] The input of the surrounding vehicle behavior intention model is the first behavior characteristic, the second behavior characteristic, the target emotion characteristic corresponding to the target vehicle driving behavior, and the road geometry characteristic, and the output of the surrounding vehicle behavior intention model is the behavior intention of the target vehicle.

[0050] In some optional embodiments, the behavior cognition-driven surrounding vehicle trajectory prediction method in a multi-vehicle interference scenario also includes a process of training a surrounding vehicle behavior intention model, and the process of training the surrounding vehicle behavior intention model includes: establishing a surrounding vehicle behavior intention model based on a first preset algorithm; inputting the acquired urban road driving data into the surrounding vehicle behavior intention model, and training the surrounding vehicle behavior intention model; wherein the input of the surrounding vehicle behavior intention model is the urban road driving data, and the output of the surrounding vehicle behavior intention model is the behavior intention information of the surrounding vehicles.

[0051] Among them, the first preset algorithm can be a conditional probability reasoning algorithm of a dynamic Bayesian network.

[0052] The embodiment of the present invention trains the surrounding vehicle behavior intention model based on urban road driving data, and continuously improves the accuracy of the surrounding vehicle behavior intention model, so that in the application of the surrounding vehicle behavior intention model, the predicted surrounding vehicle behavior intention is more accurate.

[0053] Step S105, constructing an environmental dynamic model based on the acquired road geometric features, road dynamic elements, multi-vehicle interactive interference motion characteristics and multi-vehicle interference interactive intention information; and determining target road condition information of the target vehicle based on the environmental dynamic model.

[0054] Among them, the expression of the environmental dynamic model is:

[0055] G=f d (R,D,C i ,Y);

[0056] Among them, G represents the environment dynamic model, f d () represents the dynamic environment representation function considering multi-vehicle interaction, R is the road geometry feature, D is the road dynamic element, C i represents the second behavioral characteristic, and Y represents the behavioral intention of the target vehicle.

[0057] Among them, road geometric features represent inherent attributes of roads, including road type, lane width, and road intersection type, etc. Road dynamic elements represent dynamic road traffic information obtained through the vehicle-road-cloud integrated fusion system, including road traffic flow changes, traffic signal phase information, road speed limit change information, etc.

[0058] Step S106, inputting the behavior intention and the target road condition information into the trained trajectory prediction model to obtain the trajectory prediction result of the target vehicle, wherein the target road condition information represents the dynamic condition information of the road.

[0059] In some optional implementations, the trajectory prediction result includes a speed prediction result and a turning angle prediction result of the target vehicle.

[0060] The trajectory prediction result of the embodiment of the present invention includes the speed prediction result and the turning angle prediction result of the target vehicle, and the speed prediction result and the turning angle prediction result are used to predict the longitudinal and lateral trajectory of the target vehicle, making the trajectory prediction result more accurate.

[0061] In some optional embodiments, the behavior cognition-driven surrounding vehicle trajectory prediction method in a multi-vehicle interference scenario also includes a process of training a trajectory prediction model, and the process of training the trajectory prediction model includes: constructing a trajectory prediction model based on a second preset algorithm; inputting the acquired preset behavior intention information and preset road condition information of the surrounding vehicles into the trajectory prediction model, and training the trajectory prediction model; wherein the input of the trajectory prediction model is the preset behavior intention information and the preset road condition information, and the output of the trajectory prediction model is the trajectory prediction information of the surrounding vehicles.

[0062] Among them, the second preset algorithm can be a conditional generative adversarial network algorithm, and the preset road condition information represents information data used to train the trajectory prediction model, which can be historical data or real-time training data.

[0063] The embodiment of the present invention trains the trajectory prediction model according to preset behavior intention information and preset road condition information, and continuously improves the accuracy of the trajectory prediction model, so that in the application of the trajectory prediction model, the predicted trajectory information is more accurate.

[0064] The behavior cognition-driven surrounding vehicle trajectory prediction method in the multi-vehicle interference scenario provided in this embodiment determines the target vehicle among the surrounding vehicles of the self-vehicle that has a direct impact on the safe driving of the self-vehicle in the multi-vehicle interference scenario according to the acquired urban road driving data, determines the first behavior characteristic of the self-vehicle and the second behavior characteristic of the target vehicle according to the urban road driving data, determines the target emotional characteristics corresponding to the driving behavior of the target vehicle according to the environmental information and the driver status monitoring information of the target vehicle, and obtains the behavior intention of the target vehicle according to the first behavior characteristic, the second behavior characteristic, the target emotional characteristics corresponding to the driving behavior of the target vehicle and the acquired road geometric characteristics. The embodiment of the present invention considers the influence of the driver's emotional characteristics on the driving intention, establishes the representation relationship between the driver's emotions and the movement changes of the surrounding vehicles, and more accurately infers the driving intention of the surrounding vehicles. The embodiment of the present invention takes the first behavior characteristic of the self-vehicle, the second behavior characteristic of the target vehicle, the target emotional characteristics corresponding to the driving behavior of the target vehicle, and the road geometric characteristics as the influencing factors of the behavior intention of the target vehicle, and the consideration range is wide, so the behavior intention obtained is more accurate. In the interactive process facing the multi-vehicle interference on urban roads, the embodiment of the present invention, especially the interactive behavior law of the surrounding vehicles of the multi-vehicle interference, accurately analyzes and determines the surrounding vehicle movement characteristics and driving intention of the main interference factor. The embodiment of the present invention constructs an environmental dynamic model based on the acquired road geometry features, road dynamic elements, multi-vehicle interactive interference motion characteristics, and multi-vehicle interference interactive intention information; based on the environmental dynamic model, the target road condition information of the target vehicle is determined. Many factors are considered when constructing the dynamic environmental model, so that the target road condition information is more accurate and more in line with the actual road condition information. The embodiment of the present invention inputs the behavioral intention and the target road condition information into the trained trajectory prediction model to obtain the trajectory prediction result of the target vehicle. Since the behavioral intention and the target road condition information are more accurate, the trajectory prediction result obtained based on the behavioral intention and the target road condition information is more accurate. The present invention considers the motion characteristics of the self-vehicle and the target vehicle and the emotional characteristics of the driver of the target vehicle, and significantly improves the trajectory prediction accuracy of the target vehicle of the main interference factor in the multi-vehicle interference scenario.

[0065] In this embodiment, a method for predicting surrounding vehicle trajectories driven by behavior cognition in a multi-vehicle interference scenario is provided, which can be used in the above-mentioned computer device. Figure 3 FIG. 4 is a flowchart of another method for predicting surrounding vehicle trajectories in a multi-vehicle interference scenario driven by behavior cognition according to an embodiment of the present invention. Figure 3 As shown, the process includes the following steps:

[0066] Step S301, based on the acquired urban road driving data, determine the target vehicle with main interference factors among the surrounding vehicles of the own vehicle; wherein the main interference factors represent the interference behaviors and interference vehicles that have a direct impact on the safe driving of the own vehicle in the multi-vehicle interference scenario.

[0067] Specifically, step S301 includes:

[0068] Step S3011, analyzing the multi-vehicle interference behavior of surrounding vehicles based on the urban road driving data; wherein the multi-vehicle interference behavior represents the type of impact on the driving of the own vehicle when the driving behavior and driving speed of the surrounding vehicles change.

[0069] Among them, based on urban road driving data, the multi-vehicle interference behavior in the sections, intersections and ramps of urban roads is analyzed. Multi-vehicle interference refers to the situation when the traffic flow density is large, due to the changes in the driving behavior and driving speed of surrounding vehicles (number ≥ 2, close to the safe distance range ≤ 1.5m), the interaction affects the driving of the vehicle; multi-vehicle interference behavior includes: interference from surrounding vehicles when driving on sections, interference from surrounding vehicles when driving at intersections, and interference from surrounding vehicles when merging into ramps.

[0070] Step S3012: determining the target vehicle with main interference factors among the surrounding vehicles according to the multi-vehicle interference behaviors.

[0071] Among them, the main interference factor refers to the interfering vehicles and interfering behaviors that have a direct impact on the safe driving of the vehicle in the interactive scenario of multiple vehicles. The vehicle with the main interference factor is selected as the target vehicle among the surrounding vehicles.

[0072] Step S302, determining a first behavior characteristic and a second behavior characteristic based on the urban road driving data; wherein the first behavior characteristic represents at least one of the acceleration change, speed change, direction angle, yaw angular velocity and driving intention change generated by the vehicle under the interactive interference behavior of surrounding vehicles; and the second behavior characteristic represents at least one of the acceleration change, speed change, direction angle, yaw angular velocity and driving intention change generated by the target vehicle during driving.

[0073] Specifically, step S302 includes:

[0074] Step S3021, determining a first behavior characteristic of the ego vehicle and a second behavior characteristic of the target vehicle according to the multi-vehicle interference behavior.

[0075] In an embodiment of the present invention, the first behavior characteristics of the own vehicle and the second behavior characteristics of the target vehicle are determined according to the multi-vehicle interference behavior, so as to facilitate the subsequent prediction of the intended behavior according to the first behavior characteristics of the own vehicle and the second behavior characteristics of the target vehicle.

[0076] Step S303, determining the target emotion characteristics corresponding to the driving behavior of the target vehicle according to the environmental information and the driver status monitoring information of the target vehicle.

[0077] Specifically, step S303 includes:

[0078] Step S3031, selecting a target emotional feature corresponding to the emotional state corresponding to the target vehicle driving behavior from preset emotional features based on the environmental information and the driver status monitoring information.

[0079] Among them, the preset emotional characteristics include multiple emotional states, specifically including good emotions, abnormal emotions and extreme emotions.

[0080] In the embodiment of the present invention, a representation relationship between the target emotion feature and the target vehicle motion feature change is established as follows:

[0081]

[0082] Among them, M i represents the representation relationship between the target emotion feature and the target vehicle motion feature change, f represents the representation function of the representation relationship between the target emotion feature and the target vehicle motion feature change, Indicates the driving characteristics of the target vehicle at the current moment when the driver is in a good mood. The driver is in a good mood. Indicates the driving characteristics of the target vehicle at the current moment when the driver is in an abnormal mood. The driver is in an abnormal mood. Indicates the driving characteristics of the target vehicle at the current moment when the driver is in extreme emotional state. Indicates that the driver is in extreme emotions.

[0083] The embodiment of the present invention determines the target emotional feature corresponding to the emotional state corresponding to the driving behavior of the target vehicle, and uses the target emotional feature as a factor for determining the behavioral intention of the target vehicle, which is more in line with the actual situation and the obtained behavioral intention is more accurate.

[0084] Step S304: Input the first behavior characteristic, the second behavior characteristic, the target emotion characteristic corresponding to the target vehicle driving behavior, and the acquired road geometry characteristic into the trained surrounding vehicle behavior intention model to obtain the behavior intention of the target vehicle; wherein the road geometry characteristic represents the static characteristic of the road. For details, please refer to Figure 1 Step S104 of the illustrated embodiment will not be described in detail here.

[0085] Step S305: construct an environmental dynamic model based on the acquired road geometry features, road dynamic elements, multi-vehicle interaction interference motion characteristics, and multi-vehicle interference interaction intention information; determine the target road condition information of the target vehicle based on the environmental dynamic model. Figure 1 Step S105 of the illustrated embodiment will not be described in detail here.

[0086] Step S306: Input the behavior intention and target road condition information into the trained trajectory prediction model to obtain the trajectory prediction result of the target vehicle, wherein the target road condition information represents the dynamic condition information of the road. Figure 1 Step S106 of the illustrated embodiment will not be described in detail here.

[0087] The behavior cognition-driven surrounding vehicle trajectory prediction method in a multi-vehicle interference scenario provided in this embodiment analyzes the multi-vehicle interference behavior that affects the driving of the own vehicle when the driving behavior and driving speed of the surrounding vehicles change based on urban road driving data, and determines the target vehicle with the main interference factor among the surrounding vehicles based on the multi-vehicle interference behavior. The vehicle that directly affects the driving of the own vehicle is selected from the surrounding vehicles as the target vehicle, and the target vehicle is mainly studied in the subsequent research, so that the trajectory prediction result of the target vehicle is more representative.

[0088] In this embodiment, a method for predicting surrounding vehicle trajectories driven by behavior cognition in a multi-vehicle interference scenario is provided, which can be used in the above-mentioned computer device. Figure 4 FIG. 4 is a flowchart of a method for predicting surrounding vehicle trajectories driven by behavior cognition in a multi-vehicle interference scenario according to another embodiment of the present invention. Figure 4 As shown, the process includes:

[0089] Analysis of multi-vehicle interference behavior patterns, modeling of multi-vehicle interference dynamic environments, and prediction of surrounding vehicle trajectories.

[0090] Among them, the analysis of the multi-vehicle interference behavior rules includes: multi-vehicle interference behavior analysis, obtaining multiple multi-vehicle interference behaviors, multi-vehicle interference characteristics and rules, and obtaining the behavior parameters of surrounding vehicles under multi-vehicle interference behaviors.

[0091] The multi-vehicle interference dynamic environment modeling includes: inferring the driving intentions of surrounding vehicles based on the random forest algorithm, establishing a multi-vehicle interference dynamic environment model, and obtaining target road condition information.

[0092] The trajectory prediction of surrounding vehicles includes: constructing a trajectory prediction model based on a second preset algorithm according to the cognitive intention of surrounding vehicles, inputting the driving intention and target road condition information into the trajectory prediction model, and obtaining the trajectory prediction result of the target vehicle.

[0093] In this embodiment, a behavior-recognition-driven surrounding vehicle trajectory prediction device in a multi-vehicle interference scenario is also provided, which is used to implement the above-mentioned embodiments and preferred implementation modes, and will not be repeated hereafter. As used below, the term "module" may be a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation in hardware, or a combination of software and hardware, is also possible and conceivable.

[0094] This embodiment provides a device for predicting surrounding vehicle trajectories driven by behavior cognition in a multi-vehicle interference scenario. Figure 5 As shown, including:

[0095] The target vehicle determination module 501 is used to determine the target vehicle with main interference factors among the surrounding vehicles of the own vehicle based on the acquired urban road driving data; wherein the main interference factors refer to the interference vehicles, interference behaviors and interference vehicles that have a direct impact on the safe driving of the own vehicle in the scenario of multi-vehicle interference.

[0096] The behavior characteristic determination module 502 is used to determine the first behavior characteristic and the second behavior characteristic based on the urban road driving data; wherein the first behavior characteristic represents at least one of the acceleration change, speed change, direction angle, yaw angular velocity and driving intention change generated by the vehicle under the interactive interference behavior of the surrounding vehicles; the second behavior characteristic represents at least one of the acceleration change, speed change, direction angle, yaw angular velocity and driving intention change generated by the target vehicle during driving.

[0097] The emotion feature determination module 503 is used to determine the target emotion feature corresponding to the driving behavior of the target vehicle according to the environmental information and the driver status monitoring information of the target vehicle.

[0098] The behavior intention determination module 504 is used to input the first behavior characteristic, the second behavior characteristic, the target emotional characteristic corresponding to the target vehicle driving behavior, and the acquired road geometry characteristics into the trained surrounding vehicle behavior intention model to obtain the behavior intention of the target vehicle; wherein the road geometry characteristics represent the static characteristics of the road.

[0099] The target road condition information determination module 505 is used to construct an environmental dynamic model based on the acquired road geometric features, road dynamic elements, multi-vehicle interactive interference motion characteristics and multi-vehicle interference interaction intention information; and determine the target road condition information of the target vehicle based on the environmental dynamic model.

[0100] The trajectory prediction result determination module 506 is used to input the behavior intention and the target road condition information into the trained trajectory prediction model to obtain the trajectory prediction result of the target vehicle, wherein the target road condition information represents the dynamic condition information of the road.

[0101] In some optional implementations, the target vehicle determination module 501 includes:

[0102] The multi-vehicle interference behavior determination unit is used to analyze the multi-vehicle interference behavior of surrounding vehicles based on urban road driving data; wherein the multi-vehicle interference behavior represents the type of impact on the driving of the own vehicle when the driving behavior and driving speed of the surrounding vehicles change.

[0103] The target vehicle determination unit is used to determine the target vehicle with main interference factors among the surrounding vehicles according to the interference behaviors of multiple vehicles.

[0104] In some optional implementations, the behavior characteristic determination module 502 includes:

[0105] The behavior characteristic determination unit is used to determine the first behavior characteristic of the own vehicle and the second behavior characteristic of the target vehicle according to the multi-vehicle interference behavior.

[0106] In some optional implementations, the emotion feature determination module 503 includes:

[0107] The target emotion feature selection unit is used to select a target emotion feature corresponding to the emotional state corresponding to the target vehicle driving behavior from the preset emotion features according to the environmental information and the driver status monitoring information; wherein the preset emotion features include multiple emotion states.

[0108] In some optional implementations, the device for predicting surrounding vehicle trajectories driven by behavior cognition in a multi-vehicle interference scenario further includes:

[0109] The surrounding vehicle behavior intention model training module is used to establish a surrounding vehicle behavior intention model based on a first preset algorithm; the acquired urban road driving data is input into the surrounding vehicle behavior intention model to train the surrounding vehicle behavior intention model; wherein the input of the surrounding vehicle behavior intention model is the urban road driving data, and the output of the surrounding vehicle behavior intention model is the behavior intention information of the surrounding vehicles.

[0110] The trajectory prediction model training module is used to build a trajectory prediction model based on a second preset algorithm; input the preset behavior intention information and preset road condition information of the surrounding vehicles into the trajectory prediction model to train the trajectory prediction model; wherein the input of the trajectory prediction model is the preset behavior intention information and the preset road condition information, and the output of the trajectory prediction model is the trajectory prediction information of the surrounding vehicles.

[0111] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0112] In this embodiment, the behavior cognition-driven surrounding vehicle trajectory prediction device in a multi-vehicle interference scenario is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

Claims

1. A method for predicting surrounding vehicle trajectories driven by behavior cognition in a multi-vehicle interference scenario, characterized in that: The method comprises: According to the acquired urban road driving data, a target vehicle with a main interference factor among the surrounding vehicles of the own vehicle is determined; wherein the main interference factor refers to the interference behavior and the interference vehicle that have a direct impact on the safe driving of the own vehicle in the scenario of multi-vehicle interference; Determine a first behavior characteristic and a second behavior characteristic according to the urban road driving data; wherein the first behavior characteristic represents at least one of an acceleration change, a speed change, a direction angle, a yaw rate, and a driving intention change generated by the vehicle under the interactive interference behavior of surrounding vehicles; and the second behavior characteristic represents at least one of an acceleration change, a speed change, a direction angle, a yaw rate, and a driving intention change generated by the target vehicle during driving; Determining a target emotion characteristic corresponding to the driving behavior of the target vehicle based on the environmental information and the driver status monitoring information of the target vehicle; Inputting the first behavior characteristic, the second behavior characteristic, the target emotion characteristic corresponding to the target vehicle driving behavior, and the acquired road geometry characteristic into the trained surrounding vehicle behavior intention model to obtain the behavior intention of the target vehicle; wherein the road geometry characteristic represents the static characteristic of the road; Constructing an environmental dynamic model based on the acquired road geometric features, road dynamic elements, multi-vehicle interactive interference motion characteristics, and multi-vehicle interference interaction intention information; determining target road condition information of the target vehicle based on the environmental dynamic model; Inputting the behavior intention and the target road condition information into the trained trajectory prediction model to obtain the trajectory prediction result of the target vehicle, wherein the target road condition information represents the dynamic condition information of the road; Among them, the expression of the environmental dynamic model is: in, represents the dynamic model of the environment, () represents the dynamic environment representation function considering multi-vehicle interaction, is the road geometry feature, is the dynamic element of the road, Indicates the second behavioral characteristic, Indicates the behavior intention of the target vehicle.

2. The method according to claim 1, characterized in that: The step of determining the target vehicle having the main interference factor among the surrounding vehicles according to the acquired urban road driving data includes: Analyzing the multi-vehicle interference behavior of the surrounding vehicles according to the urban road driving data; wherein the multi-vehicle interference behavior represents the type of influence on the driving of the own vehicle when the driving behavior and driving speed of the surrounding vehicles change; According to the multi-vehicle interference behaviors, a target vehicle having a main interference factor among the surrounding vehicles is determined.

3. The method according to claim 2, characterized in that The determining the first behavior characteristic and the second behavior characteristic according to the urban road driving data includes: The first behavior characteristic of the ego vehicle and the second behavior characteristic of the target vehicle are determined according to the multi-vehicle interference behavior.

4. The method according to claim 2 or 3, characterized in that: The multi-vehicle interference behavior includes the interference of surrounding vehicles when driving on a road section, the interference of surrounding vehicles when driving at an intersection, and the interference of surrounding vehicles when merging into a ramp.

5. The method according to any one of claims 1 to 3, characterized in that The step of determining the target emotion feature corresponding to the target vehicle driving behavior according to the environmental information and the driver status monitoring information includes: According to the environmental information and the driver status monitoring information, a target emotional feature corresponding to the emotional state corresponding to the target vehicle driving behavior is selected from the preset emotional features; wherein the preset emotional features include multiple emotional states.

6. The method according to any one of claims 1 to 3, characterized in that The trajectory prediction result includes a vehicle speed prediction result and a turning angle prediction result of the target vehicle.

7. The method according to any one of claims 1 to 3, characterized in that The method also includes a process of training the surrounding vehicle behavior intention model, and the process of training the surrounding vehicle behavior intention model includes: Based on a first preset algorithm, establishing the surrounding vehicle behavior intention model; The acquired urban road driving data is input into the surrounding vehicle behavior intention model to train the surrounding vehicle behavior intention model; wherein the input of the surrounding vehicle behavior intention model is the urban road driving data, and the output of the surrounding vehicle behavior intention model is the behavior intention information of the surrounding vehicles.

8. The method according to any one of claims 1 to 3, characterized in that The method also includes a process of training the trajectory prediction model, and the process of training the trajectory prediction model includes: Based on a second preset algorithm, construct the trajectory prediction model; The acquired preset behavior intention information and preset road condition information of surrounding vehicles are input into the trajectory prediction model to train the trajectory prediction model; wherein the input of the trajectory prediction model is the preset behavior intention information and the preset road condition information, and the output of the trajectory prediction model is the trajectory prediction information of surrounding vehicles.

9. A device for predicting surrounding vehicle trajectories driven by behavior cognition in a multi-vehicle interference scenario, characterized in that: The device comprises: A target vehicle determination module is used to determine a target vehicle with a main interference factor among the surrounding vehicles of the vehicle according to the acquired urban road driving data; wherein the main interference factor refers to an interference vehicle, interference behavior and interference vehicle that have a direct impact on the safe driving of the vehicle in a multi-vehicle interference scenario; A behavior characteristic determination module, for determining a first behavior characteristic and a second behavior characteristic according to the urban road driving data; wherein the first behavior characteristic represents at least one of an acceleration change, a speed change, a direction angle, a yaw angular velocity, and a driving intention change generated by the vehicle under the interactive interference behavior of surrounding vehicles; and the second behavior characteristic represents at least one of an acceleration change, a speed change, a direction angle, a yaw angular velocity, and a driving intention change generated by the target vehicle during driving; An emotion feature determination module, used to determine a target emotion feature corresponding to the driving behavior of the target vehicle according to environmental information and driver status monitoring information of the target vehicle; A behavior intention determination module, used to input the first behavior characteristic, the second behavior characteristic, the target emotional characteristic corresponding to the driving behavior of the target vehicle, and the acquired road geometric characteristics into the trained surrounding vehicle behavior intention model to obtain the behavior intention of the target vehicle; wherein the road geometric characteristics represent static characteristics of the road; The target road condition information determination module is used to construct an environmental dynamic model based on the acquired road geometric features, road dynamic elements, multi-vehicle interactive interference motion characteristics and multi-vehicle interference interaction intention information; and determine the target road condition information of the target vehicle based on the environmental dynamic model; A trajectory prediction result determination module, used for inputting the behavior intention and the target road condition information into the trained trajectory prediction model to obtain the trajectory prediction result of the target vehicle, wherein the target road condition information represents the dynamic condition information of the road; Among them, the expression of the environmental dynamic model is: in, represents the dynamic model of the environment, () represents the dynamic environment representation function considering multi-vehicle interaction, is the road geometry feature, is the dynamic element of the road, Indicates the second behavioral characteristic, Indicates the behavior intention of the target vehicle.

Images