WO2023045265A1 - 人-车-路微观交通系统建模及风险辨识方法及装置 - Google Patents

人-车-路微观交通系统建模及风险辨识方法及装置 Download PDF

Info

Publication number
WO2023045265A1
WO2023045265A1 PCT/CN2022/079109 CN2022079109W WO2023045265A1 WO 2023045265 A1 WO2023045265 A1 WO 2023045265A1 CN 2022079109 W CN2022079109 W CN 2022079109W WO 2023045265 A1 WO2023045265 A1 WO 2023045265A1
Authority
WO
WIPO (PCT)
Prior art keywords
vehicle
driver
road
interaction
human
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2022/079109
Other languages
English (en)
French (fr)
Inventor
黄荷叶
王建强
杨奕彬
刘艺璁
崔明阳
许庆
李克强
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tsinghua University
Original Assignee
Tsinghua University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Tsinghua University filed Critical Tsinghua University
Priority to JP2024516526A priority Critical patent/JP7717409B2/ja
Publication of WO2023045265A1 publication Critical patent/WO2023045265A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0635—Risk analysis of enterprise or organisation activities
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/40—Business processes related to the transportation industry
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2111/00—Details relating to CAD techniques
    • G06F2111/08—Probabilistic or stochastic CAD

Definitions

  • the present application relates to the technical field of automatic driving, and in particular to a method and device for modeling and risk identification of human-vehicle-road micro-traffic system.
  • the current research on the vehicle-road coupling mechanism is mostly based on tire dynamics-road It is difficult to expand and build a closed-loop system model of pedestrians, vehicles and roads from a unified perspective.
  • the research on human-vehicle interaction mostly focuses on the technical overcoming of advanced driving assistance systems, with vehicle safety as the main goal. Few systems consider the driver's acceptance of the assistance system, and there is a limitation of poor amenity. Understand the driver's driving cognitive characteristics, build its cognitive model, break through the trust problem in the human-vehicle interaction process, and realize human-vehicle collaborative driving.
  • the research on the human-road interaction mechanism mainly focuses on the research on the driving behavior intention of a single vehicle, and the research on the identification of the interaction intention is not yet in-depth. Therefore, it is necessary to build a human-road interaction mechanism model to find out the interaction mechanism between the driver of the vehicle and the surrounding traffic participants. , to provide theoretical support for active control to avoid collision risk.
  • the research on the closed-loop system of people-vehicle-road is mostly concentrated in the scope of narrow concepts, and some factors are oversimplified. Although it covers the elements of human-vehicle-road traffic, it lacks a unified and complete system description. Therefore, there is an urgent need for a new and scalable unified modeling method for pedestrians and vehicles to describe the complex system, and on this basis to grasp the influence of traffic environment changes on vehicle operation safety, and provide traffic managers or vehicle motion control. theoretical guidance.
  • Road traffic risks are considered to be the result of a combination of many factors, mainly manifested as drivers' physical and psychological limitations, limited vehicle performance, improper road routing, and risks caused by bad weather (such as low visibility and slippery roads).
  • a single space-time distance parameter actual vehicle distance, inter-vehicle time, and collision time, etc.
  • This application provides a human-vehicle-road micro-traffic system modeling and risk identification method to solve the problem of unclear modeling of the human-vehicle-road traffic system itself and the difficulty of effectively feeding back the results of risk identification to people in related technologies -Vehicle-road traffic system for safety auxiliary decision-making and other issues.
  • the embodiment of the first aspect of the present application provides a human-vehicle-road micro-traffic system modeling and risk identification method, including the following steps: using the interaction between different types of vehicles and road traffic participants to construct vehicle-road dynamic interaction
  • the interaction model is used to obtain the potential accident consequences caused by the vehicle-road interaction
  • the human-vehicle dynamic interaction model is constructed by using the driving trajectory distribution output by the human-vehicle system interaction, and the behavioral uncertainty caused by the human-vehicle interaction is obtained
  • the driver’s visual characteristics are used to analyze the driving Observe the interest area of the road surrounding environment during human driving to construct a human-road dynamic interaction model, and obtain the difference in driver risk sensitivity during the human-road interaction process
  • use the clustering idea to characterize the characteristic rules and differences of drivers' driving habits, Obtain the personalized characteristics of the driver; construct a human-vehicle - A closed-loop dynamic system that generates risk identification results.
  • the human-vehicle-road closed-loop dynamics system is expressed as:
  • S DVR is the human-vehicle-road closed-loop dynamics system
  • ⁇ i (t) is the effective response in the difference of driver risk sensitivity in the human-road interaction process
  • D i is the driver's individual characteristics
  • ⁇ ( t) is the steering angle of the vehicle
  • P x,y *( ⁇ (t)) is the behavioral uncertainty
  • r' is the radius of the driving area of interest in the driver's risk sensitivity difference during the human-road interaction process
  • ⁇ is steering
  • ⁇ >0 means turning left
  • ⁇ 0 means turning right
  • ⁇ is the increase of steering angle within a certain time period ⁇ t quantity
  • IN is the interaction matrix between self-vehicle i and road traffic participant j
  • m i , m j and v i , v j are the mass and speed of self-vehicle i and another road traffic participant j respectively
  • TTC is Coll
  • the vehicle-road dynamic interaction model is constructed using the interaction between different types of vehicles and road traffic participants to obtain the potential accident consequences generated by vehicle-road interaction, Including: calculating the number of interactions between the self-vehicle and different road traffic participants in the scene; calculating the interaction consequences between two objects in the traffic environment; based on the interaction quantity and the interaction consequences, generating In the process of vehicle interaction, the cumulative consequences of the interaction are superimposed to obtain the potential accident consequences generated by the vehicle-road interaction
  • IN is the interaction matrix between self-vehicle i and road traffic participant j
  • m i , m j and vi , v j are the mass and speed of self-vehicle i and another road traffic participant j, respectively.
  • the construction of a human-vehicle dynamic interaction model using the driving trajectory distribution output by the human-vehicle system interaction to obtain the behavioral uncertainty generated by the human-vehicle interaction includes: according to the vehicle Calculate the equivalent linear two-wheeled vehicle model based on the kinematics model and turning radius, and calculate the predicted position according to the command steering angle; predict the uncertain motion of the driver-vehicle system based on the collected actual driving experiment data; based on the predicted position and the uncertainty movement, obtain the Gaussian normal distribution of the steering angle of the vehicle steering angle, and determine the parameters based on the Gaussian normal distribution to determine the behavioral uncertainty P x generated by the human-vehicle interaction, y ( ⁇ (t)):
  • ⁇ (t) is the steering angle of the vehicle
  • ⁇ is the degree of dispersion of the data distribution that obeys the normal distribution
  • ⁇ is the mean value of the random variable that obeys the normal distribution.
  • the driver's visual characteristics are used to construct a human-road dynamic interaction model for the region of interest in which the driver observes the surrounding road environment during driving, and the driver's risk risk during the human-road interaction process is obtained.
  • Sensitivity differences including:
  • the equipotential line length l ⁇ of the dynamic visual range of the driver is obtained by using the oval distribution characteristics of the driver's vision:
  • r' is the radius of the area of interest
  • a is the length of the major axis of the ellipse
  • b is the length of the minor axis of the ellipse
  • the values of a and b are related to the speed v i (t) of the ego vehicle i, is the viewing angle function of the driver;
  • TTC is the collision avoidance time
  • ⁇ (v i ) a/b
  • v i (t) is the speed of ego vehicle i
  • using the clustering idea to characterize the characteristic rules and differences of the driver's driving habits to obtain the personalized characteristics of the driver includes: distinguishing the drivers of each type in the same scene Under the control behavior and track similarity, the driver's personalized representation is obtained based on the unsupervised clustering method; the characteristic representation of each driver is obtained based on the driver's personalized representation, and the driving behavior is obtained based on the distance from the cluster center. characteristics to determine the driver's individual characteristics.
  • the embodiment of the second aspect of the present application provides a human-vehicle-road micro-traffic system modeling and risk identification device, including: a first modeling module, which is used to utilize the interaction between different types of vehicles and road traffic participants Construct a vehicle-road dynamic interaction model to obtain potential accident consequences caused by vehicle-road interaction; the second modeling module is used to construct a human-vehicle dynamic interaction model using the driving trajectory distribution output by the human-vehicle system interaction to obtain human-vehicle interaction Generated behavioral uncertainty; the third modeling module is used to use the driver's visual characteristics to construct a human-road dynamic interaction model for the area of interest observed by the driver in the surrounding environment of the road during driving, and obtain the driver's risk during the human-road interaction process.
  • a first modeling module which is used to utilize the interaction between different types of vehicles and road traffic participants Construct a vehicle-road dynamic interaction model to obtain potential accident consequences caused by vehicle-road interaction
  • the second modeling module is used to construct a human-vehicle dynamic interaction model using the driving trajectory
  • Sensitivity difference used to characterize the characteristic rules and differences of the driver's driving habits by using clustering thinking, and obtain the driver's personalized characteristics
  • identification module used to generate potential accident consequences based on the vehicle-road interaction , the behavioral uncertainty, the driver's risk sensitivity difference during the human-road interaction process, and the driver's personalized characteristics construct a human-vehicle-road closed-loop dynamics system to generate risk identification results.
  • the human-vehicle-road closed-loop dynamics system is expressed as:
  • S DVR is the human-vehicle-road closed-loop dynamics system
  • ⁇ i (t) is the effective response in the difference of driver risk sensitivity in the human-road interaction process
  • D i is the driver's individual characteristics
  • ⁇ ( t) is the steering angle of the vehicle
  • P x,y *( ⁇ (t)) is the behavioral uncertainty
  • r' is the radius of the driving area of interest in the driver's risk sensitivity difference during the human-road interaction process
  • ⁇ is steering
  • ⁇ >0 means turning left
  • ⁇ 0 means turning right
  • ⁇ is the increase of steering angle within a certain time period ⁇ t quantity
  • IN is the interaction matrix between self-vehicle i and road traffic participant j
  • m i , m j and v i , v j are the mass and speed of self-vehicle i and another road traffic participant j respectively
  • TTC is Coll
  • the embodiment of the third aspect of the present application provides an electronic device, including: a memory, a processor, and a computer program stored on the memory and operable on the processor, and the processor executes the program to realize The method for modeling and risk identification of the human-vehicle-road micro-traffic system as described in the above-mentioned embodiments.
  • the embodiment of the fourth aspect of the present application provides a computer-readable storage medium, on which a computer program is stored, and the program is executed by a processor to realize the construction of the human-vehicle-road micro-traffic system as described in the above-mentioned embodiment. Model and risk identification method.
  • This application analyzes the original attributes and interaction characteristics of people, vehicles, roads, and the environment, and constructs a unified mathematical model of the human-vehicle-road micro-traffic system, which can be input to the traffic system as a whole for state analysis.
  • This application constructs a model to characterize the influence of driver factors, vehicle motion state, and road environment information interaction process on the system safety state, reveals the mechanism of micro-traffic system risk generation, and can realize system risk identification and graded early warning, and further protect the road safety of people and vehicles.
  • the transportation system is intrinsically safe.
  • this application is more effective in scenarios with arbitrary road topology (crossroads, ring roads, expressways, etc.) , can identify the dangerous situation of the human-vehicle-road micro-traffic system, and assist the vehicle in the decision-making of driving risk classification. It helps to give timely warnings or auxiliary corrections to drivers who are approaching dangerous states in complex traffic environments, and provides new ideas for the research of anti-collision warning strategies and control methods.
  • Fig. 1 is a flow chart of a method for modeling and risk identification of a human-vehicle-road micro-traffic system provided according to an embodiment of the present application;
  • FIG. 2 is a Gaussian distribution diagram of steering angle angles in a highway scene according to an embodiment of the present application
  • Fig. 3 is a division diagram of a driving operation gaze area provided according to an embodiment of the present application.
  • Fig. 4 is a schematic framework diagram of a human-vehicle-environment element coupling relationship provided according to an embodiment of the present application
  • FIG. 5 is a schematic diagram of risk evolution of a pedestrian-vehicle road system model in a multi-lane lane change and obstacle avoidance scenario provided according to an embodiment of the present application;
  • FIG. 6 is an example diagram of a person-vehicle-road micro-traffic system modeling and risk identification device according to an embodiment of the present application
  • Fig. 7 is a schematic structural diagram of an electronic device provided in an embodiment of the application.
  • first modeling module-100 second modeling module-200, third modeling module-300, characterization module-400, identification module-500, memory-701, processor-702, communication interface -703.
  • FIG. 1 is a flowchart of a method for modeling and risk identification of a human-vehicle-road micro-traffic system according to an embodiment of the present application.
  • the human-vehicle-road micro-traffic system modeling and risk identification method includes the following steps:
  • step S101 a vehicle-road dynamic interaction model is constructed using the interaction between different types of vehicles and road traffic participants to obtain potential accident consequences caused by vehicle-road interaction.
  • a vehicle-road dynamic interaction model is constructed using the interaction between different types of vehicles and road traffic participants to obtain potential accident consequences generated by vehicle-road interaction, including: calculating The number of interactions between the self-vehicle and different road traffic participants in the scene; calculate the interaction consequences between two objects in the traffic environment; based on the interaction quantity and interaction consequences, generate superimposed interactions during the interaction between multiple traffic participants and the self-vehicle The cumulative consequences obtain the potential accident consequences generated by the vehicle-road interaction.
  • E i represents the transfer energy of ego vehicle i
  • m i , m j and v i , v j are the mass and velocity (vector) of ego vehicle i and another entity j, respectively.
  • step S102 the human-vehicle dynamic interaction model is constructed using the driving trajectory distribution output by the human-vehicle system interaction, and the behavioral uncertainty caused by the human-vehicle interaction is obtained.
  • the human-vehicle dynamic interaction model is constructed using the driving trajectory distribution output by the human-vehicle system interaction, and the behavioral uncertainty generated by the human-vehicle interaction is obtained, including: according to the vehicle kinematics model Calculate the equivalent linear two-wheeled vehicle model with the turning radius, and calculate the predicted position according to the command steering angle; predict the uncertain motion of the driver-vehicle system based on the collected actual driving experiment data; predict the uncertain motion based on the predicted position and uncertain motion , the Gaussian normal distribution of the steering angle of the vehicle steering angle is obtained, and the parameters are determined based on the Gaussian normal distribution to determine the behavioral uncertainty caused by human-vehicle interaction.
  • the vehicle in the process of human-vehicle interaction, the vehicle is equivalent to an actuator with feedback, which can be stimulated by the uncertainty of the driver's intention and the dynamically adjusted driving target, and output the dynamic response of the vehicle during driving, then
  • the interaction output of the human-vehicle system can be described by the distribution of driving trajectories, and the behavioral uncertainty P x,y ( ⁇ (t)) generated by the human-vehicle interaction can be output.
  • the turning radius R can be calculated with the equivalent linear two-wheeled vehicle model as follows:
  • K is the stability coefficient
  • L is the wheelbase of vehicle i
  • ⁇ (t) is the steering angle of the vehicle.
  • the predicted position can be calculated as:
  • the possible trajectory of vehicle i should have a certain boundary, and the motion state of vehicle i is absolutely stable within this boundary.
  • the driver may drive straight, turn to the left lane, and turn to the right lane.
  • the uncertain motion of the driver-vehicle system can be predicted based on the actual driving experiment data.
  • the statistics of the steering angle of the vehicle show that the distribution of the steering angle basically presents a Gaussian normal distribution, and the parameters of the Gaussian distribution can be determined based on the experimental data:
  • ⁇ (t) is the steering angle of the vehicle
  • ⁇ is the degree of dispersion of the data distribution that obeys the normal distribution
  • ⁇ is the mean value of the random variable that obeys the normal distribution.
  • the specific parameters of the Gaussian distribution may be slightly different due to the differences in the characteristics of the vehicle being controlled by different drivers, but the overall trend is consistent, that is, the steering response amplitude of the driver's control of the vehicle shows a Gaussian distribution.
  • step S103 the driver's visual characteristics are used to construct a human-road dynamic interaction model for the region of interest observed by the driver in the surrounding environment of the road during driving, and the difference in risk sensitivity of the driver during the human-road interaction process is obtained.
  • the driver's visual characteristics are used to construct a human-road dynamic interaction model for the region of interest in which the driver observes the surrounding road environment during driving, and the driver's risk sensitivity during the human-road interaction process is obtained
  • the differences include: using the driver's visual range to calculate the effective response of the driver during normal driving; using the driver's visual oval distribution characteristics to obtain the equipotential line length of the driver's dynamic visual range; The impact is used to perceive the relative speed of the vehicle, and the influence of the driver's dynamic perception field of view is obtained, and the radius of the area of interest size for observing the surrounding environment of the road is generated during driving, and the difference in risk sensitivity of the driver during the human-road interaction process is determined according to the effective response and the radius of the size of the area of interest .
  • the driver mainly relies on vision to obtain information during driving, so the interaction between the driver and the road environment is mainly affected by the driver's visual effect.
  • the driver's visual characteristics model the interest area of the driver's observation of the surrounding environment of the road during driving, and output the driver's risk sensitivity difference during the human-road interaction process (including effective response ⁇ i (t), driving area of interest radius r′) .
  • the driver's response to the surrounding environment changes with his viewing angle and scanning frequency, and he is most sensitive to the relative distance and relative speed of other vehicles in the road environment from himself.
  • the front of the driver is the driver's keen vision area, and the driver must always pay attention and respond in time.
  • On both sides of the sharp vision area and the front area of the adjacent lane is the driver's general vision area, but he often pays attention to it during driving, because the car in the adjacent lane may cut in, and he may also need to change to the adjacent lane. .
  • the areas on both sides, namely the peripheral visual area, are rarely paid attention to, and the rear of the vehicle is usually ignored due to factors such as visual characteristics and responsibility identification.
  • ⁇ i which is the angle of view function, where is the driver's perspective.
  • l is the distance from the point on the equipotential line to car i
  • a is the length of the major axis of the ellipse
  • b is the length of the minor axis of the ellipse
  • ⁇ v is the relative speed of the vehicle
  • D p is the distance between vehicles
  • TTC is the collision avoidance time (s)
  • v i is the driving speed of vehicle i.
  • TTC is the collision avoidance time
  • ⁇ (v i ) a/b
  • v i (t) the speed of ego vehicle i, which can define the radius of the area of interest for the driver to observe the surrounding environment of the road during driving as r '.
  • step S104 the characteristic rules and differences of the driver's driving habits are characterized by using the clustering idea to obtain the personalized characteristics of the driver.
  • the clustering idea is used to characterize the characteristic rules and differences of the driving habits of the drivers to obtain the personalized characteristics of the drivers, including: Manipulation behavior and trajectory similarity, based on the unsupervised clustering method to obtain the driver's personalized representation; based on the driver's personalized representation to obtain the characteristic representation of each driver, based on the distance from the cluster center to obtain the driving characteristics, determine the driving personal characteristics of people.
  • the clustering idea is used to characterize the characteristic rules and differences of the driver's driving habits, and the driver's personalized characteristic D i is output.
  • step S105 construct a human-vehicle-road closed-loop dynamics system based on potential accident consequences generated by vehicle-road interaction, behavioral uncertainty, driver risk sensitivity differences during human-road interaction, and driver individual characteristics, and generate risk identification result.
  • the human-vehicle-road closed-loop dynamics system is expressed as:
  • S DVR is the human-vehicle-road closed-loop dynamic system
  • ⁇ i (t) is the effective response to the difference of driver risk sensitivity in the process of human-road interaction
  • D i is the driver's individual characteristics
  • ⁇ (t) is the vehicle steering Angle
  • P x,y *( ⁇ (t)) is the behavior uncertainty
  • r′ is the radius of the driving area of interest in the driver’s risk sensitivity difference during the human-road interaction process
  • S DVR is the human-vehicle-road closed-loop dynamic system
  • ⁇ i (t) is the effective response to the difference of driver risk sensitivity in the process of human-road interaction
  • D i is the driver's individual characteristics
  • ⁇ (t) is the vehicle steering Angle
  • P x,y *( ⁇ (t)) is the behavior uncertainty
  • r′ is the radius of the driving area of interest in the driver’s risk sensitivity difference during the human-road interaction process
  • ⁇ is steering
  • ⁇ >0 means turning left
  • ⁇ 0 means turning right
  • ⁇ is the increment of steering angle within a certain period of time ⁇ t
  • IN is the interaction matrix between self-vehicle i and road traffic participant j
  • m i , m j and v i , v j are the mass and speed of self-vehicle i and another road traffic participant j respectively
  • TTC collision avoidance
  • the driving environment of the car is complex, the operating states are diverse, and the driving style and driving experience of the driver are different.
  • the driver, vehicle, and road environment are coupled to each other to form a complex and generalized dynamical system, and the safety of the system is affected by the interaction between the driver, the vehicle, and the road.
  • the human-vehicle-road closed-loop system has strong nonlinear, coupling and time-varying characteristics.
  • the vehicle is equivalent to a control system, and the movement of the vehicle is like the operation of the controller.
  • the driver observes the surrounding environment, such as road conditions, objects visible in the field of vision, etc., and sends corresponding signals to the vehicle, that is, the operation signals applied to the steering wheel, brake, and accelerator pedal. Subsequently, the driver receives the status feedback signal from the vehicle through sensory organs such as eyes and ears, and adjusts the control accordingly.
  • the environment also affects the driver and the vehicle. In poor road conditions, the vehicle may not respond accurately to control operations, or the driver's driving behavior may change. So the environment is part of the noise and disturbances that affect the vehicle control system.
  • the controller By adding a controller to the vehicle, it can make adjustments for situations where the vehicle malfunctions or becomes difficult to control. Through environmental input, the controller understands the driving environment and road conditions, and through driver input, the controller understands the driver's operating status. The controller is usually also required to provide signals to the vehicle to adjust the driver's behavior according to the vehicle's own driving conditions, such as steering angle, braking or traction torque, etc. Therefore, the entire pedestrian-vehicle road system can be represented as shown in Figure 4.
  • the driving risk belongs to the generalized instability of the human-vehicle-road closed-loop system, which is usually jointly determined by the driver, road conditions, and the dynamic characteristics of the vehicle itself.
  • the driver drives the vehicle normally on a straight road with good road conditions.
  • the system state is SDVR,0 .
  • the accelerator pedal will be released or the brake pedal will be pressed to decelerate the vehicle so as to maintain a relatively safe distance from the obstacle vehicle in front.
  • the state of the human-vehicle-road system evolves from SDVR,1 to the state of S DVR,2 , when the driver makes a sudden mistake, the human-vehicle-road system
  • the state changes from S DVR,2 to S DVR,3 , and an accident occurs.
  • the human-vehicle-road micro-traffic system modeling and risk identification method proposed according to the embodiment of the application is used to characterize the influence of driver factors, vehicle motion state, and road environment information interaction process on the system safety state, and reveals the micro-traffic system. Risk generation mechanism to realize system risk identification and graded early warning.
  • the human-vehicle-road unified model analyzes the original attributes and interaction characteristics of people, vehicles, roads, and environments.
  • the road micro-traffic system identifies dangerous situations and assists vehicles in grading decision-making for driving risks. This method helps to give timely warnings or auxiliary corrections to drivers who are approaching dangerous states in complex traffic environments, and provides a new idea for the research of anti-collision warning strategies and control methods.
  • FIG. 6 is an example diagram of a human-vehicle-road micro-traffic system modeling and risk identification device according to an embodiment of the present application.
  • the person-vehicle-road micro-traffic system modeling and risk identification device 10 includes: a first modeling module 100, a second modeling module 200, a third modeling module 300, a characterization module 400 and an identification Module 500.
  • the first modeling module 100 is used to construct a vehicle-road dynamic interaction model by utilizing the interaction between different types of vehicles and road traffic participants, and obtain potential accident consequences caused by vehicle-road interaction.
  • the second modeling module 200 is configured to construct a human-vehicle dynamic interaction model using the driving trajectory distribution output by the human-vehicle system interaction, and obtain behavioral uncertainty generated by the human-vehicle interaction.
  • the third modeling module 300 is used to use the driver's visual characteristics to construct a human-road dynamic interaction model for the interest area observed by the driver in the surrounding environment of the road during driving, and obtain the difference in risk sensitivity of the driver during the human-road interaction process.
  • the characterization module 400 is used to characterize the characteristic rules and differences of the driver's driving habits by using the clustering idea to obtain the driver's personalized characteristics.
  • the identification module 500 is used to construct a human-vehicle-road closed-loop dynamics system based on potential accident consequences generated by vehicle-road interaction, behavioral uncertainty, driver risk sensitivity differences during human-road interaction, and driver individual characteristics to generate risk Identification results.
  • the human-vehicle-road closed-loop dynamics system is expressed as:
  • S DVR is the human-vehicle-road closed-loop dynamic system
  • ⁇ i (t) is the effective response to the difference of driver risk sensitivity in the process of human-road interaction
  • D i is the driver's individual characteristics
  • ⁇ (t) is the vehicle steering Angle
  • P x,y *( ⁇ (t)) is the behavior uncertainty
  • r′ is the radius of the driving area of interest in the driver’s risk sensitivity difference during the human-road interaction process
  • ⁇ is steering
  • ⁇ >0 means turning left
  • ⁇ 0 means turning right
  • ⁇ is the increment of steering angle within a certain period of time ⁇ t
  • IN is the interaction matrix between self-vehicle i and road traffic participant j
  • m i , m j and v i , v j are the mass and speed of self-vehicle i and another road traffic participant j respectively
  • TTC collision avoidance
  • the first modeling module is specifically used to calculate the number of interactions between the ego vehicle and different road traffic participants in the scene; calculate the interaction consequences between two objects in the traffic environment; Based on the number of interactions and the consequences of interactions, the potential accident consequences of vehicle-road interactions are generated by superimposing the cumulative consequences of interactions in the process of interaction between multiple traffic participants and the self-vehicle
  • IN is the interaction matrix between self-vehicle i and road traffic participant j
  • m i , m j and vi , v j are the mass and speed of self-vehicle i and another road traffic participant j, respectively.
  • the second modeling module is specifically configured to calculate an equivalent linear two-wheeled vehicle model according to the vehicle kinematics model and the turning radius, and calculate the predicted position according to the command steering angle; based on The collected actual driving experiment data predicts the uncertain movement of the driver-vehicle system; based on the predicted position and uncertain movement, the Gaussian normal distribution of the steering angle of the vehicle steering angle is obtained, and the parameters are calculated based on the Gaussian normal distribution. Determine, determine the behavioral uncertainty P x,y ( ⁇ (t)) generated by human-vehicle interaction:
  • ⁇ (t) is the steering angle of the vehicle
  • ⁇ is the degree of dispersion of the data distribution that obeys the normal distribution
  • ⁇ is the mean value of the random variable that obeys the normal distribution.
  • the third modeling module is specifically configured to use the driver's visual range to calculate the driver's effective response during normal driving;
  • the equipotential line length l ⁇ of the dynamic visual range of the driver is obtained by using the oval distribution characteristics of the driver's vision:
  • r' is the radius of the area of interest
  • a is the length of the major axis of the ellipse
  • b is the length of the minor axis of the ellipse
  • the values of a and b are related to the speed v i (t) of the ego vehicle i, is the viewing angle function of the driver;
  • TTC is the collision avoidance time
  • ⁇ (v i ) a/b
  • v i (t) is the speed of ego vehicle i
  • the difference in the driver's risk sensitivity during the human-road interaction process is determined.
  • the characterization module is specifically used to distinguish the steering behavior and trajectory similarity of each type of driver in the same scene, and obtain the personalized characterization of the driver based on the unsupervised clustering method ;
  • the characteristic representation of each driver is obtained based on the personalized representation of the driver, the driving characteristics are obtained based on the distance from the cluster center, and the personalized characteristics of the driver are determined.
  • the human-vehicle-road micro-traffic system modeling and risk identification device proposed according to the embodiment of the application is used to characterize the influence of driver factors, vehicle motion state, and road environment information interaction process on the system safety state, and reveals the micro-traffic system. Risk generation mechanism to realize system risk identification and graded early warning.
  • the human-vehicle-road unified model analyzes the original attributes and interaction characteristics of people, vehicles, roads, and environments.
  • the road micro-traffic system identifies dangerous situations and assists vehicles in grading decision-making for driving risks. It helps to give timely warnings or auxiliary corrections to drivers who are approaching dangerous states in complex traffic environments, and provides new ideas for the research of anti-collision warning strategies and control methods.
  • FIG. 7 is a schematic structural diagram of an electronic device provided by an embodiment of the present application.
  • This electronic equipment can include:
  • a memory 701 a processor 702 , and a computer program stored in the memory 701 and executable on the processor 702 .
  • processor 702 executes the program, it implements the human-vehicle-road micro-traffic system modeling and risk identification methods provided in the above-mentioned embodiments.
  • the electronic equipment also includes:
  • the communication interface 703 is used for communication between the memory 701 and the processor 702 .
  • the memory 701 is used to store computer programs that can run on the processor 702 .
  • the memory 701 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.
  • the bus may be an Industry Standard Architecture (ISA) bus, a Peripheral Component (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus.
  • ISA Industry Standard Architecture
  • PCI Peripheral Component
  • EISA Extended Industry Standard Architecture
  • the bus can be divided into address bus, data bus, control bus and so on. For ease of representation, only one thick line is used in FIG. 7 , but it does not mean that there is only one bus or one type of bus.
  • the memory 701, processor 702, and communication interface 703 are integrated on one chip, then the memory 701, processor 702, and communication interface 703 can communicate with each other through the internal interface.
  • the processor 702 may be a central processing unit (Central Processing Unit, referred to as CPU), or a specific integrated circuit (Application Specific Integrated Circuit, referred to as ASIC), or be configured to implement one or more of the embodiments of the present application integrated circuit.
  • CPU Central Processing Unit
  • ASIC Application Specific Integrated Circuit
  • This embodiment also provides a computer-readable storage medium, on which a computer program is stored, which is characterized in that, when the program is executed by a processor, the above method for modeling and risk identification of the human-vehicle-road micro-traffic system is realized.
  • first and second are used for descriptive purposes only, and cannot be interpreted as indicating or implying relative importance or implicitly specifying the quantity of indicated technical features.
  • the features defined as “first” and “second” may explicitly or implicitly include at least one of these features.
  • “N” means at least two, such as two, three, etc., unless otherwise specifically defined.
  • Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code comprising one or more executable instructions for implementing a custom logical function or step of a process , and the scope of preferred embodiments of the present application includes additional implementations in which functions may be performed out of the order shown or discussed, including in substantially simultaneous fashion or in reverse order depending on the functions involved, which shall It should be understood by those skilled in the art to which the embodiments of the present application belong.
  • each part of the present application may be realized by hardware, software, firmware or a combination thereof.
  • the N steps or methods may be implemented by software or firmware stored in memory and executed by a suitable instruction execution system.
  • a suitable instruction execution system For example, if implemented in hardware as in another embodiment, it can be implemented by any one or a combination of the following techniques known in the art: a discrete Logic circuits, ASICs with suitable combinational logic gates, Programmable Gate Arrays (PGA), Field Programmable Gate Arrays (FPGA), etc.

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • General Physics & Mathematics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Marketing (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Primary Health Care (AREA)
  • Quality & Reliability (AREA)
  • Development Economics (AREA)
  • Educational Administration (AREA)
  • Health & Medical Sciences (AREA)
  • Game Theory and Decision Science (AREA)
  • Operations Research (AREA)
  • General Health & Medical Sciences (AREA)
  • Computer Hardware Design (AREA)
  • Evolutionary Computation (AREA)
  • Geometry (AREA)
  • General Engineering & Computer Science (AREA)
  • Traffic Control Systems (AREA)
  • Control Of Driving Devices And Active Controlling Of Vehicle (AREA)

Abstract

本申请公开了一种人-车-路微观交通系统建模及风险辨识方法及装置，其中，方法包括：构建车-路动态交互作用模型，得到车路交互产生的潜在事故后果；构建人-车动态交互作用模型，得到人车交互产生的行为不确定性；构建人-路动态交互作用模型，得到人路交互过程驾驶人风险敏感度差异；利用聚类思想将驾驶人驾驶习性的特征规律和差异性进行表征，得到驾驶人个性化特性；根据车路交互产生的潜在事故后果、行为不确定性、人路交互过程驾驶人风险敏感度差异及驾驶人个性化特性构建人-车-路闭环动力学系统，生成风险辨识结果。由此，可以表征驾驶人因素、车辆运动状态、道路环境信息交互过程对系统安全状态的影响，实现系统风险辨识与分级预警。

Description

人-车-路微观交通系统建模及风险辨识方法及装置
相关申请的交叉引用
本申请要求清华大学于2021年09月26日提交的、发明名称为“人-车-路微观交通系统建模及风险辨识方法及装置”的、中国专利申请号“202111131805.6”的优先权。
技术领域
本申请涉及自动驾驶技术领域，特别涉及一种人-车-路微观交通系统建模及风险辨识方法及装置。
背景技术
车辆行驶时，道路环境复杂多变，其安全状态受到较多因素影响，大体上可以分为人(生理状态、操纵特征、意图等)、车(运行工况)、路(天气、路况、交通流等)三个因素。由于驾驶人预瞄不足、反应延迟以及操控失误等不确定性因素，加之可能出现的车辆故障、周车行为意图难以预测以及道路环境恶化等，使得交通事故难以从根本上彻底消除。此外，驾驶人特性、周车意图和道路环境的不确定性随着人-车-路闭环系统非线性迭代不断传播放大，导致车辆动力学特性难以预测。因此，阐明人-车-路交互作用，准确表征各交通要素对系统的影响，量化辨识复杂系统动态风险，可以更好适应驾驶人、车辆和道路环境的不同条件，从而平衡人-车-路微观交通系统。
相关技术中，人-车-路交互作用，尤其是人-车、车-路、人-路交互之间建模已经获得了广泛的研究。驾驶人(人)、车辆(车)、交通环境(路)三要素及其相互作用，构成了以自车为中心的微观交通系统基本单元。现有研究多从单一对象扩展至人-车、人-路、车-路交互视角，即较少从人-车-路统一视角出发。具体来说，现有的车-路耦合研究愈发深入，但针对极限工况下的车-路交互耦合机理研究尚不清晰，当前关于车-路耦合机理的研究多基于轮胎动力学-道路动力学的作用机制来展开，难扩展构建统一视角下人车路闭环系统模型。而人车交互的研究多聚焦面向先进驾驶辅助系统的技术攻克，以车辆安全性为主要目标，鲜有系统考虑驾驶人对辅助系统的接受程度，存在宜人性较差的局限，势必需要通过深刻认识驾驶人驾驶认知特性，构建其认知模型，突破人车交互过程的信任问题，实现人车协同驾驶。而人-路交互机理研究主要集中在单一车辆驾驶行为意图研究，对交互意图辨识研究尚不深入，因此需要构建人-路交互机理模型，探明自车驾驶人与周围交通参与者的交互机制，为规避碰撞风险的主动控制提供理论支撑。对于人-车-路闭环系统的研究多集中在狭义概念范围内，对某些因素过度简化，虽涵盖人车路交通要素，但缺乏统一、完整的系统 描述。因此，迫切需要一种全新可扩展的人车路统一建模方法，对该复杂系统展开描述，在此基础上掌握交通环境变化对车辆运行安全的影响规律，为交通管理者或车辆运动控制提供理论指导。
道路交通风险产生被认为是许多因素综合的结果，主要表现为驾驶人的生理和心理限制、车辆性能有限、道路路线不当以及恶劣天气造成的风险(如能见度低和道路湿滑)。针对人-车-路闭环系统这一复杂耦合对象，从行车风险形成开始到发生危险冲突的整个风险转化过程很难用单一的时空距离参数(实际车距、车间时间和碰撞时间等)进行描述，需要综合考虑多个时空距离参数并采用更复杂的模型和算法对系统运行风险进行研究。但现有基于车辆运动学、碰撞概率等方法对于人-车-路交通系统本身的建模不清晰，难将风险辨识的结果有效反馈于人-车-路交通系统进行安全辅助决策。因此，在复杂人-车-路交通系统中，需要充分考虑系统中多源多维的风险产生过程，辨识人-车-路耦合环境下的车辆行驶安全状态，构建人-车-路闭环系统风险辨识模型，进而更好实现人-车-路闭环系统运行本质安全。因此，有必要开发一种人-车-路微观交通系统建模及风险辨识方法。
发明内容
本申请提供一种人-车-路微观交通系统建模及风险辨识方法，以解决相关技术中对于人-车-路交通系统本身的建模不清晰，难将风险辨识的结果有效反馈于人-车-路交通系统进行安全辅助决策等问题。
本申请第一方面实施例提供一种人-车-路微观交通系统建模及风险辨识方法，包括以下步骤：利用不同类型的车与道路交通参与者之间的交互作用构建车-路动态交互作用模型，得到车路交互产生的潜在事故后果；利用人车系统交互输出的行驶轨迹分布构建人-车动态交互作用模型，得到人车交互产生的行为不确定性；利用驾驶人视觉特性对驾驶人驾驶过程中观察道路周边环境的兴趣区域构建人-路动态交互作用模型，得到人路交互过程驾驶人风险敏感度差异；利用聚类思想将驾驶人驾驶习性的特征规律和差异性进行表征，得到驾驶人个性化特性；根据所述车路交互产生的潜在事故后果、所述行为不确定性、所述人路交互过程驾驶人风险敏感度差异及所述驾驶人个性化特性构建人-车-路闭环动力学系统，生成风险辨识结果。
可选地，在本申请的一个实施例中，所述人-车-路闭环动力学系统表示为：
Figure PCTCN2022079109-appb-000001
其中，S DVR为人-车-路闭环动力学系统，γ i(t)为所述人路交互过程驾驶人风险敏感度差异中的有效响应，D i为所述驾驶人个性化特性，δ(t)为车辆转向角度，P x,y*(δ(t))为所 述行为不确定性，r′为所述人路交互过程驾驶人风险敏感度差异中的驾驶兴趣区域半径，
Figure PCTCN2022079109-appb-000002
为所述车路交互产生的潜在事故后果，ω为转向，ω＝0为车辆直行，ω>0代表左转，ω<0代表右转，Δδ为在某一时间段Δt内转向角的增量，IN为自车i与道路交通参与者j进行交互的交互矩阵，m i,m j和v i,v j分别是自车i与另一个道路交通参与者j的质量和速度，TTC是避撞时间，ξ(v i)＝a/b，a为椭圆长轴长度，b为椭圆短轴长度，v i(t)为自车i的速度，
Figure PCTCN2022079109-appb-000003
为驾驶人的视角函数。
为可选地，在本申请的一个实施例中，所述利用不同类型的车与道路交通参与者之间的交互作用构建车-路动态交互作用模型，得到车路交互产生的潜在事故后果，包括：计算场景中自车与不同道路交通参与者的交互数量；计算交通环境中两个对象之间的交互后果；基于所述交互数量和所述交互后果，生成在多个交通参与者与自车进行交互过程中叠加交互累计后果得到所述车路交互产生的潜在事故后果
Figure PCTCN2022079109-appb-000004
Figure PCTCN2022079109-appb-000005
其中，IN为自车i与道路交通参与者j进行交互的交互矩阵，m i,m j和v i,v j分别是自车i与另一个道路交通参与者j的质量和速度。
为可选地，在本申请的一个实施例中，所述利用人车系统交互输出的行驶轨迹分布构建人-车动态交互作用模型，得到人车交互产生的行为不确定性，包括：根据车辆运动学模型和转弯半径计算等效线性两轮车模型，并根据指令转向角计算预测位置；基于采集的实际驾驶实验数据对驾驶人-车辆系统的不确定性运动进行预测；基于所述预测位置和所述不确定性运动，得到所述车辆转向角的转角角度的高斯正态分布，并基于所述高斯正态分布进行参数确定，确定所述人车交互产生的行为不确定性P x,y(δ(t))：
Figure PCTCN2022079109-appb-000006
其中，δ(t)为车辆转向角度，σ为服从正态分布的数据分布的离散程度，μ为服从正态分布的随机变量的均值。
可选地，在本申请的一个实施例中，所述利用驾驶人视觉特性对驾驶人驾驶过程中观察道路周边环境的兴趣区域构建人-路动态交互作用模型，得到人路交互过程驾驶人风险敏感度差异，包括：
利用驾驶人的视觉范围计算驾驶人正常驾驶过程中的有效响应；
利用驾驶人视觉椭圆形分布特性获取驾驶人动态视觉范围的等势线长度lρ：
Figure PCTCN2022079109-appb-000007
其中，r′为兴趣区域大小的半径，a为椭圆长轴长度，b为椭圆短轴长度，a和b取值与自车i车速度v i(t)相关，
Figure PCTCN2022079109-appb-000008
为驾驶人的视角函数；
利用驾驶人视角变化对驾驶人的影响来感知车辆相对速度，得到驾驶人动态感知视野影响，生成驾驶过程中观察道路周边环境的兴趣区域大小的半径：
Figure PCTCN2022079109-appb-000009
其中，TTC是避撞时间，ξ(v i)＝a/b，v i(t)为自车i的速度；
根据所述有效响应以及所述兴趣区域大小的半径确定所述人路交互过程驾驶人风险敏感度差异。
可选地，在本申请的一个实施例中，所述利用聚类思想将驾驶人驾驶习性的特征规律和差异性进行表征，得到驾驶人个性化特性，包括：区分每类驾驶人在同一场景下的操纵行为和轨迹相似度，基于非监督聚类方法得到驾驶人的个性化表征；基于所述驾驶人的个性化表征得到每个驾驶人的特性表征，基于距离聚类中心的距离得到驾驶特性，确定所述驾驶人个性化特性。
本申请第二方面实施例提供一种人-车-路微观交通系统建模及风险辨识装置，包括：第一建模模块，用于利用不同类型的车与道路交通参与者之间的交互作用构建车-路动态交互作用模型，得到车路交互产生的潜在事故后果；第二建模模块，用于利用人车系统交互输出的行驶轨迹分布构建人-车动态交互作用模型，得到人车交互产生的行为不确定性；第三建模模块，用于利用驾驶人视觉特性对驾驶人驾驶过程中观察道路周边环境的兴趣区域构建人-路动态交互作用模型，得到人路交互过程驾驶人风险敏感度差异；表征模块，用于利用聚类思想将驾驶人驾驶习性的特征规律和差异性进行表征，得到驾驶人个性化特性；辨识模块，用于根据所述车路交互产生的潜在事故后果、所述行为不确定性、所述人路交互过程驾驶人风险敏感度差异及所述驾驶人个性化特性构建人-车-路闭环动力学系统，生成风险辨识结果。
可选地，在本申请的一个实施例中，所述人-车-路闭环动力学系统表示为：
Figure PCTCN2022079109-appb-000010
其中，S DVR为人-车-路闭环动力学系统，γ i(t)为所述人路交互过程驾驶人风险敏感度差异中的有效响应，D i为所述驾驶人个性化特性，δ(t)为车辆转向角度，P x,y*(δ(t))为所述行为不确定性，r′为所述人路交互过程驾驶人风险敏感度差异中的驾驶兴趣区域半径，
Figure PCTCN2022079109-appb-000011
为 所述车路交互产生的潜在事故后果，ω为转向，ω＝0为车辆直行，ω>0代表左转，ω<0代表右转，Δδ为在某一时间段Δt内转向角的增量，IN为自车i与道路交通参与者j进行交互的交互矩阵，m i,m j和v i,v j分别是自车i与另一个道路交通参与者j的质量和速度，TTC是避撞时间，ξ(v i)＝a/b，a为椭圆长轴长度，b为椭圆短轴长度，v i(t)为自车i的速度，
Figure PCTCN2022079109-appb-000012
为驾驶人的视角函数。
本申请第三方面实施例提供一种电子设备，包括：存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机程序，所述处理器执行所述程序，以实现如上述实施例所述的人-车-路微观交通系统建模及风险辨识方法。
本申请第四方面实施例提供一种计算机可读存储介质，其上存储有计算机程序，该程序被处理器执行，以用于实现如上述实施例所述的人-车-路微观交通系统建模及风险辨识方法。
本申请实施例的人-车-路微观交通系统建模及风险辨识方法及装置，具有以下优势：
1)本申请分析了人、车、路及环境等本源属性和交互特性，构建人-车-路微观交通系统的统一数学模型，能作为整体输入到交通系统进行状态分析。
2)本申请构建模型表征驾驶人因素、车辆运动状态、道路环境信息交互过程对系统安全状态的影响，揭示了微观交通系统风险产生机理，可以实现系统风险辨识与分级预警，进一步保障人车路交通系统本质安全。
3)本申请相比避撞时间(Time To Collision，TTC)、车头时距(Time Headway，THW)等其他风险指标，在具有任意道路拓扑的场景(十字路口、环形路、高速公路等)中，都可以对人-车-路微观交通系统进行危险态势辨识，并辅助车辆进行驾驶风险分级决策。有助于在复杂的交通环境下对临近危险状态的驾驶人给予及时的警告或辅助纠正，为防撞预警策略和控制方法的研究提供了新的思路。
本申请附加的方面和优点将在下面的描述中部分给出，部分将从下面的描述中变得明显，或通过本申请的实践了解到。
附图说明
本申请上述的和/或附加的方面和优点从下面结合附图对实施例的描述中将变得明显和容易理解，其中：
图1为根据本申请实施例提供的一种人-车-路微观交通系统建模及风险辨识方法的流程图；
图2为根据本申请实施例提供的一种高速公路场景下转向角角度高斯分布图；
图3为根据本申请实施例提供的一种驾驶操作注视区域划分图；
图4为根据本申请实施例提供的一种人-车-环境要素耦合关系示意性框架图；
图5为根据本申请实施例提供的一种多车道换道避障场景下人车路系统模型风险演变示意图；
图6为根据本申请实施例的人-车-路微观交通系统建模及风险辨识装置的示例图；
图7为申请实施例提供的电子设备的结构示意图。
附图标记说明：第一建模模块-100、第二建模模块-200、第三建模模块-300、表征模块-400、辨识模块-500、存储器-701、处理器-702、通信接口-703。
具体实施方式
下面详细描述本申请的实施例，所述实施例的示例在附图中示出，其中自始至终相同或类似的标号表示相同或类似的元件或具有相同或类似功能的元件。下面通过参考附图描述的实施例是示例性的，旨在用于解释本申请，而不能理解为对本申请的限制。
具体而言，图1为根据本申请实施例提供的一种人-车-路微观交通系统建模及风险辨识方法的流程图。
如图1所示，该人-车-路微观交通系统建模及风险辨识方法包括以下步骤：
在步骤S101中，利用不同类型的车与道路交通参与者之间的交互作用构建车-路动态交互作用模型，得到车路交互产生的潜在事故后果。
可选地，在本申请的一个实施例中，利用不同类型的车与道路交通参与者之间的交互作用构建车-路动态交互作用模型，得到车路交互产生的潜在事故后果，包括：计算场景中自车与不同道路交通参与者的交互数量；计算交通环境中两个对象之间的交互后果；基于交互数量和交互后果，生成在多个交通参与者与自车进行交互过程中叠加交互累计后果得到车路交互产生的潜在事故后果。
具体地，人车路微观交通系统运行过程中，自车和道路交通参与者之间会存在多类型交互行为，为表征交互过程产生的可能后果，针对不同类型的车与道路交通参与者之间的交互作用进行建模，获取车路交互产生的潜在事故后果
Figure PCTCN2022079109-appb-000013
首先，计算场景中自车与不同道路交通参与者的交互数量。在任意一个场景下，场景中包含不同道路交通参与者J＝[j 0,j 1,j 2,…j n]，自车i与参与者j进行交互可以用交互矩阵IN来表征，
Figure PCTCN2022079109-appb-000014
其中in 0,0＝1表示自车与其他道路交通参与者j进行交互，而in 0,0＝0表示不交互。
其次，计算交通环境中两个对象之间的交互后果；考虑交互过程如不进行及时防控则会出现碰撞事故，产生实体之间能量转移，即在发生碰撞的情况下，产生的后果可以看作 是碰撞实体之间的能量传递，即：
Figure PCTCN2022079109-appb-000015
其中，E i表示自车i的传递能量，m i,m j和v i,v j分别是自车i与另一个实体j的质量和速度(矢量)。
最后，在多个交通参与者与自车进行交互过程中，即将交互累计后果进行叠加得到车路交互产生的潜在事故后果：
Figure PCTCN2022079109-appb-000016
在步骤S102中，利用人车系统交互输出的行驶轨迹分布构建人-车动态交互作用模型，得到人车交互产生的行为不确定性。
可选地，在本申请的一个实施例中，利用人车系统交互输出的行驶轨迹分布构建人-车动态交互作用模型，得到人车交互产生的行为不确定性，包括：根据车辆运动学模型和转弯半径计算等效线性两轮车模型，并根据指令转向角计算预测位置；基于采集的实际驾驶实验数据对驾驶人-车辆系统的不确定性运动进行预测；基于预测位置和不确定性运动，得到车辆转向角的转角角度的高斯正态分布，并基于高斯正态分布进行参数确定，确定人车交互产生的行为不确定性。
具体地，在人和车辆交互过程中，车辆相当于一个的带反馈的执行机构，其能够受到驾驶人意图的不确定性和动态调整的驾驶目标的激励，输出车辆的行驶过程动态响应，则可将人车系统交互输出用行驶轨迹分布来进行描述，输出人车交互产生的行为不确定性P x,y(δ(t))。
首先，假设正常驾驶人驾驶遵守交通规则和法律。因此，车辆正常行驶时主要进行直行、左转、右转操纵行为，并伴随速度的动态调整。根据车辆运动学模型，转弯半径R可以用等效线性两轮车模型计算如下：
Figure PCTCN2022079109-appb-000017
其中，K为稳定系数，L为车辆i的轴距，δ(t)为车辆转向角度。
当车辆i以可忽略侧滑角的恒定速度行驶时，在指令转向角δ下，预测位置
Figure PCTCN2022079109-appb-000018
可计算为：
Figure PCTCN2022079109-appb-000019
其次，根据转向角范围，车辆i的可能运动轨迹应具有一定的边界，并且车辆i的运动 状态绝对稳定在该边界内。当车辆i在道路上直线行驶时，驾驶人可能进行直线行驶，转向左侧车道，转向右侧车道。为确定驾驶人操纵车辆后，车辆输出的位置范围，可基于实际驾驶实验数据，对驾驶人-车辆系统的不确定性运动进行预测。
最后，统计车辆转向角度得出转角角度分布基本呈现高斯正态分布，并可基于实验数据对高斯分布进行参数确定：
Figure PCTCN2022079109-appb-000020
其中，δ(t)为车辆转向角度，σ为服从正态分布的数据分布的离散程度，μ为服从正态分布的随机变量的均值。
参见图2，本申请实施例提供的高速公路场景下转向角角度高斯分布图所示，假设以北京市一段环行路的实际数据采集结果作为输入，可以拟合得到高斯分布的具体参数得：μ＝0,σ＝3.77，则输出转角角度分布为：
Figure PCTCN2022079109-appb-000021
在不同环境下，由于不同驾驶人操控车辆特性的差异性，会导致高斯分布的具体参数有一点差异，但整体趋势具有一致性，即驾驶人操控车辆过程的转向响应幅度呈高斯分布。
在步骤S103中，利用驾驶人视觉特性对驾驶人驾驶过程中观察道路周边环境的兴趣区域构建人-路动态交互作用模型，得到人路交互过程驾驶人风险敏感度差异。
可选地，在本申请的一个实施例中，利用驾驶人视觉特性对驾驶人驾驶过程中观察道路周边环境的兴趣区域构建人-路动态交互作用模型，得到人路交互过程驾驶人风险敏感度差异，包括：利用驾驶人的视觉范围计算驾驶人正常驾驶过程中的有效响应；利用驾驶人视觉椭圆形分布特性获取驾驶人动态视觉范围的等势线长度；利用驾驶人视角变化对驾驶人的影响来感知车辆相对速度，得到驾驶人动态感知视野影响，生成驾驶过程中观察道路周边环境的兴趣区域大小的半径，根据有效响应以及兴趣区域大小的半径确定人路交互过程驾驶人风险敏感度差异。
具体地，驾驶人在驾驶过程中主要依靠视觉获取信息，因此驾驶人与道路环境的交互主要受驾驶人视觉效应的影响。根据驾驶人视觉特性对驾驶人驾驶过程中观察道路周边环境的兴趣区域进行建模，输出人路交互过程驾驶人风险敏感度差异(包括有效响应γ i(t)、驾驶兴趣区域半径r′)。
首先，计算驾驶人正常驾驶过程中的有效响应。驾驶人对于周围环境的响应随着其视角和扫描频率而变化，并对道路环境中其他车辆与其自身的相对距离和相对速度的感知最为敏锐。驾驶员正前方位为驾驶员的敏锐视觉区、驾驶员必须经常关注并且及时做出反应。 而敏锐视觉区两边、相邻车道的前方区域是驾驶员的一般视觉区，但在驾驶过程中也时常关注，这是因为邻道的车可能插进来，同时自己也可能需要换到邻道。而两侧的区域，即外围视觉区则很少关注，车辆的后方由于视觉特性和责任认定等因素通常会被忽略。假设驾驶员对于周围环境做出的反应为γ i，是其视角
Figure PCTCN2022079109-appb-000022
的函数，其中
Figure PCTCN2022079109-appb-000023
是驾驶员的视角。比如，取驾驶人有效响应γ i(0)＝1,γ i(π)＝0,，那么当正前方出现状况时，驾驶员对于做出100％的响应，而对于正后方尾随的车辆则处于忽略状态。
其次，参见图3，本申请实施例提供的驾驶操作注视区域划分图所示，根据驾驶人视觉椭圆形分布特性，假设驾驶人处于椭圆的一个焦点处F 1处，而驾驶人动态视野与车速成反比，车速越快视野也越窄，忽略正后方的视觉区域，可以得出在
Figure PCTCN2022079109-appb-000024
的动态视觉范围的等势线长度：
Figure PCTCN2022079109-appb-000025
式中，
Figure PCTCN2022079109-appb-000026
为驾驶人的视角函数，l为等势线上点距离i车的距离，a为椭圆长轴长度，b为椭圆短轴长度，其取值与i车速度v i(t)相关。即速度越快，视线范围变窄。
再次，根据驾驶人视觉特性，视角的变化对驾驶人的影响从而可以感知车辆相对速度，进而影响驾驶人动态感知视野，即影响lρ。视角变化率
Figure PCTCN2022079109-appb-000027
与lρ有如下关系：
Figure PCTCN2022079109-appb-000028
其中，Δv为车辆的相对速度；D p为车辆间距，TTC是避撞时间(s)，v i是车辆i行驶速度。
将上式进行变化，定义ξ(v i)＝a/b，可求得等效半径为：
Figure PCTCN2022079109-appb-000029
其中，TTC是避撞时间，ξ(v i)＝a/b，v i(t)为自车i的速度，即可定义驾驶人驾驶过程中观察道路周边环境的兴趣区域大小的半径为r′。
在步骤S104中，利用聚类思想将驾驶人驾驶习性的特征规律和差异性进行表征，得到驾驶人个性化特性。
可选地，在本申请的一个实施例中，利用聚类思想将驾驶人驾驶习性的特征规律和差异性进行表征，得到驾驶人个性化特性，包括：区分每类驾驶人在同一场景下的操纵行为和轨迹相似度，基于非监督聚类方法得到驾驶人的个性化表征；基于驾驶人的个性化表征得到每个驾驶人的特性表征，基于距离聚类中心的距离得到驾驶特性，确定驾驶人个性化 特性。
具体地，由于驾驶人个体的差异性及工况的复杂性，在真实交通场景下驾驶人主动行为呈现出高度随机性和非线性等特征。因此，利用聚类思想将驾驶人驾驶习性的特征规律和差异性进行表征，输出驾驶人个性化特性D i。
首先，不同年龄、不同性别、不同性格、不同驾龄和不同驾驶熟练程度等都会对驾驶习性产生显著影响，驾驶人的个性化表征对系统稳定性密切相关。按照每类驾驶员在同一场景下的操纵行为和轨迹相似度进行区分，基于非监督聚类方法输出3个聚类中心(C1 x,y,C2 x,y,C3 x,y)，将驾驶员分为3类，定义为激进型A、正常型B和保守型C。
其次，进行个性化驾驶人特性表征，可以表示为该驾驶人行为(如操纵轨迹)再进行聚类Ci x,y，分别得出距离3个聚类中心的距离(|Ci x,y-C1 x,y|，|Ci x,y-C2 x,y|，|Ci x,y-C3 x,y|)，并计算距离比例为：
Figure PCTCN2022079109-appb-000030
Figure PCTCN2022079109-appb-000031
Figure PCTCN2022079109-appb-000032
即可获得该驾驶人的驾驶特性为(D i＝k 1A+k 2B+k 3C)。
在步骤S105中，根据车路交互产生的潜在事故后果、行为不确定性、人路交互过程驾驶人风险敏感度差异及驾驶人个性化特性构建人-车-路闭环动力学系统，生成风险辨识结果。
可选地，在本申请的一个实施例中，人-车-路闭环动力学系统表示为：
Figure PCTCN2022079109-appb-000033
其中，S DVR为人-车-路闭环动力学系统，γ i(t)为人路交互过程驾驶人风险敏感度差异中的有效响应，D i为驾驶人个性化特性，δ(t)为车辆转向角度，P x,y*(δ(t))为行为不确定性，r′为人路交互过程驾驶人风险敏感度差异中的驾驶兴趣区域半径，
Figure PCTCN2022079109-appb-000034
为车路交互产生的潜在事故后果。
具体地，通过分别对车-路、人-车、人-路的交互作用进行建模，即通过车路交互产生的潜在事故后果
Figure PCTCN2022079109-appb-000035
人车交互产生的行为不确定性P x,y(δ(t))和人路交互过程驾驶人风险敏感度差异(包括有效响应γ i(t)、驾驶兴趣区域半径r′，驾驶人个性化特性D i)，最终以统一 视角构建人-车-路闭环动力学系统。系统统一模型S DVR表示如下：
Figure PCTCN2022079109-appb-000036
其中，S DVR为人-车-路闭环动力学系统，γ i(t)为人路交互过程驾驶人风险敏感度差异中的有效响应，D i为驾驶人个性化特性，δ(t)为车辆转向角度，P x,y*(δ(t))为行为不确定性，r′为人路交互过程驾驶人风险敏感度差异中的驾驶兴趣区域半径，
Figure PCTCN2022079109-appb-000037
为车路交互产生的潜在事故后果，ω为转向，ω＝0为车辆直行，ω>0代表左转，ω<0代表右转，Δδ为在某一时间段Δt内转向角的增量，IN为自车i与道路交通参与者j进行交互的交互矩阵，m i,m j和v i,v j分别是自车i与另一个道路交通参与者j的质量和速度，TTC是避撞时间，ξ(v i)＝a/b，a为椭圆长轴长度，b为椭圆短轴长度，v i(t)为自车i的速度，
Figure PCTCN2022079109-appb-000038
为驾驶人的视角函数。
参见图4，采用本申请实施例提供的人-车-环境要素耦合关系示意性框架图所示，复杂工况下，汽车驾驶环境复杂，运行状态多样，驾驶员驾驶风格和驾驶体验不同，驾驶员、车辆、道路环境相互耦合形成一个复杂、广义的动力学系统，系统安全性受驾驶员交互作用的影响，车辆和道路。人-车-路闭环系统具有很强的非线性、耦合和时变特性。
从整体上看，车辆相当于一个控制系统，车辆的运动就像控制器的操作。驾驶车辆时，驾驶员观察周围的环境，如路况、视野内可见的物体等，并向车辆发送相应的信号，即应用于方向盘、刹车、油门踏板的操作信号。随后，驾驶员通过眼、耳等感觉器官接收到来自车辆的状态反馈信号，并据此调整控制。同时，环境也会影响驾驶员和车辆。如果路况较差，车辆可能无法准确响应控制操作，或者驾驶员的驾驶行为可能发生变化。所以环境是影响车辆控制系统的噪声和干扰的一部分。
通过在车辆上添加控制器，它可以对车辆故障或难以控制的情况进行调整。通过环境输入，控制器了解驾驶环境和道路状况，通过驾驶员输入，控制器了解驾驶员的运行状态。通常还要求控制器根据车辆自身的驾驶状况，如转向角度、制动或牵引力矩等，向车辆提供信号，以调整驾驶员的行为。因此，整个人车路系统可以表示为如图4所示。
如图5所示，行车风险属于人-车-路闭环系统的广义失稳，通常由驾驶人、道路条件、以及车辆本身的动力学特性共同决定。以多车道换道避障场景为例，初始时刻，驾驶人驾驶车辆在路况良好平直路段正常驾驶，此阶段系统状态为S DVR,0，当车辆行驶车道前方出现障碍，驾驶人发现前方障碍后将会开始松开加速踏板或踩下制动踏板实现车辆减速，以保持和前方障碍车的相对安全距离，人-车-路系统状态变化为S DVR,1，随后驾驶人持续减速过 程中，车速不断降低，自车与障碍车之间距离不断缩短，人-车-路系统状态由S DVR,1演变为状态S DVR,2，当驾驶人出现突然失误时，人-车-路系统状态由S DVR,2演变为状态S DVR,3，事故发生。
根据本申请实施例提出的人-车-路微观交通系统建模及风险辨识方法，用于表征驾驶人因素、车辆运动状态、道路环境信息交互过程对系统安全状态的影响，揭示了微观交通系统风险产生机理，实现系统风险辨识与分级预警。该人-车-路统一模型分析了人、车、路及环境等本源属性和交互特性，在具有任意道路拓扑的场景(十字路口、环形路、高速公路等)中，都可以对人-车-路微观交通系统进行危险态势辨识，并辅助车辆进行驾驶风险分级决策。该方法有助于在复杂的交通环境下对临近危险状态的驾驶人给予及时的警告或辅助纠正，为防撞预警策略和控制方法的研究提供了新的思路。
其次参照附图描述根据本申请实施例提出的人-车-路微观交通系统建模及风险辨识装置。
图6为根据本申请实施例的人-车-路微观交通系统建模及风险辨识装置的示例图。
如图6所示，该人-车-路微观交通系统建模及风险辨识装置10包括：第一建模模块100、第二建模模块200、第三建模模块300、表征模块400和辨识模块500。
其中，第一建模模块100，用于利用不同类型的车与道路交通参与者之间的交互作用构建车-路动态交互作用模型，得到车路交互产生的潜在事故后果。第二建模模块200，用于利用人车系统交互输出的行驶轨迹分布构建人-车动态交互作用模型，得到人车交互产生的行为不确定性。第三建模模块300，用于利用驾驶人视觉特性对驾驶人驾驶过程中观察道路周边环境的兴趣区域构建人-路动态交互作用模型，得到人路交互过程驾驶人风险敏感度差异。表征模块400，用于利用聚类思想将驾驶人驾驶习性的特征规律和差异性进行表征，得到驾驶人个性化特性。辨识模块500，用于根据车路交互产生的潜在事故后果、行为不确定性、人路交互过程驾驶人风险敏感度差异及驾驶人个性化特性构建人-车-路闭环动力学系统，生成风险辨识结果。
可选地，在本申请的一个实施例中，人-车-路闭环动力学系统表示为：
Figure PCTCN2022079109-appb-000039
其中，S DVR为人-车-路闭环动力学系统，γ i(t)为人路交互过程驾驶人风险敏感度差异中的有效响应，D i为驾驶人个性化特性，δ(t)为车辆转向角度，P x,y*(δ(t))为行为不确定性，r′为人路交互过程驾驶人风险敏感度差异中的驾驶兴趣区域半径，
Figure PCTCN2022079109-appb-000040
为车路交互产生的潜在 事故后果，ω为转向，ω＝0为车辆直行，ω>0代表左转，ω<0代表右转，Δδ为在某一时间段Δt内转向角的增量，IN为自车i与道路交通参与者j进行交互的交互矩阵，m i,m j和v i,v j分别是自车i与另一个道路交通参与者j的质量和速度，TTC是避撞时间，ξ(v i)＝a/b，a为椭圆长轴长度，b为椭圆短轴长度，v i(t)为自车i的速度，
Figure PCTCN2022079109-appb-000041
为驾驶人的视角函数。
可选地，在本申请的一个实施例中，第一建模模块，具体用于计算场景中自车与不同道路交通参与者的交互数量；计算交通环境中两个对象之间的交互后果；基于交互数量和交互后果，生成在多个交通参与者与自车进行交互过程中叠加交互累计后果得到车路交互产生的潜在事故后果
Figure PCTCN2022079109-appb-000042
Figure PCTCN2022079109-appb-000043
其中，IN为自车i与道路交通参与者j进行交互的交互矩阵，m i,m j和v i,v j分别是自车i与另一个道路交通参与者j的质量和速度。
可选地，在本申请的一个实施例中，第二建模模块，具体用于，根据车辆运动学模型和转弯半径计算等效线性两轮车模型，并根据指令转向角计算预测位置；基于采集的实际驾驶实验数据对驾驶人-车辆系统的不确定性运动进行预测；基于预测位置和不确定性运动，得到车辆转向角的转角角度的高斯正态分布，并基于高斯正态分布进行参数确定，确定人车交互产生的行为不确定性P x,y(δ(t))：
Figure PCTCN2022079109-appb-000044
其中，δ(t)为车辆转向角度，σ为服从正态分布的数据分布的离散程度，μ为服从正态分布的随机变量的均值。
可选地，在本申请的一个实施例中，第三建模模块，具体用于，利用驾驶人的视觉范围计算驾驶人正常驾驶过程中的有效响应；
利用驾驶人视觉椭圆形分布特性获取驾驶人动态视觉范围的等势线长度lρ：
Figure PCTCN2022079109-appb-000045
其中，r′为兴趣区域大小的半径，a为椭圆长轴长度，b为椭圆短轴长度，a和b取值与自车i车速度v i(t)相关，
Figure PCTCN2022079109-appb-000046
为驾驶人的视角函数；
利用驾驶人视角变化对驾驶人的影响来感知车辆相对速度，得到驾驶人动态感知视野影响，生成驾驶过程中观察道路周边环境的兴趣区域大小的半径：
Figure PCTCN2022079109-appb-000047
其中，TTC是避撞时间，ξ(v i)＝a/b，v i(t)为自车i的速度；
根据有效响应以及兴趣区域大小的半径确定人路交互过程驾驶人风险敏感度差异。
可选地，在本申请的一个实施例中，表征模块，具体用于，区分每类驾驶人在同一场景下的操纵行为和轨迹相似度，基于非监督聚类方法得到驾驶人的个性化表征；基于驾驶人的个性化表征得到每个驾驶人的特性表征，基于距离聚类中心的距离得到驾驶特性，确定驾驶人个性化特性。
需要说明的是，前述对人-车-路微观交通系统建模及风险辨识实施例的解释说明也适用于该实施例的人-车-路微观交通系统建模及风险辨识装置，此处不再赘述。
根据本申请实施例提出的人-车-路微观交通系统建模及风险辨识装置，用于表征驾驶人因素、车辆运动状态、道路环境信息交互过程对系统安全状态的影响，揭示了微观交通系统风险产生机理，实现系统风险辨识与分级预警。该人-车-路统一模型分析了人、车、路及环境等本源属性和交互特性，在具有任意道路拓扑的场景(十字路口、环形路、高速公路等)中，都可以对人-车-路微观交通系统进行危险态势辨识，并辅助车辆进行驾驶风险分级决策。有助于在复杂的交通环境下对临近危险状态的驾驶人给予及时的警告或辅助纠正，为防撞预警策略和控制方法的研究提供了新的思路。
图7为本申请实施例提供的电子设备的结构示意图。该电子设备可以包括：
存储器701、处理器702及存储在存储器701上并可在处理器702上运行的计算机程序。
处理器702执行程序时实现上述实施例中提供的人-车-路微观交通系统建模及风险辨识方法。
进一步地，电子设备还包括：
通信接口703，用于存储器701和处理器702之间的通信。
存储器701，用于存放可在处理器702上运行的计算机程序。
存储器701可能包含高速RAM存储器，也可能还包括非易失性存储器(non-volatile memory)，例如至少一个磁盘存储器。
如果存储器701、处理器702和通信接口703独立实现，则通信接口703、存储器701和处理器702可以通过总线相互连接并完成相互间的通信。总线可以是工业标准体系结构(Industry Standard Architecture，简称为ISA)总线、外部设备互连(Peripheral Component，简称为PCI)总线或扩展工业标准体系结构(Extended Industry Standard Architecture，简称为EISA)总线等。总线可以分为地址总线、数据总线、控制总线等。为便于表示，图7中仅用一条粗线表示，但并不表示仅有一根总线或一种类型的总线。
可选的，在具体实现上，如果存储器701、处理器702及通信接口703，集成在一块芯 片上实现，则存储器701、处理器702及通信接口703可以通过内部接口完成相互间的通信。
处理器702可能是一个中央处理器(Central Processing Unit，简称为CPU)，或者是特定集成电路(Application Specific Integrated Circuit，简称为ASIC)，或者是被配置成实施本申请实施例的一个或多个集成电路。
本实施例还提供一种计算机可读存储介质，其上存储有计算机程序，其特征在于，该程序被处理器执行时实现如上的人-车-路微观交通系统建模及风险辨识方法。
在本说明书的描述中，参考术语“一个实施例”、“一些实施例”、“示例”、“具体示例”、或“一些示例”等的描述意指结合该实施例或示例描述的具体特征、结构、材料或者特点包含于本申请的至少一个实施例或示例中。在本说明书中，对上述术语的示意性表述不必须针对的是相同的实施例或示例。而且，描述的具体特征、结构、材料或者特点可以在任一个或N个实施例或示例中以合适的方式结合。此外，在不相互矛盾的情况下，本领域的技术人员可以将本说明书中描述的不同实施例或示例以及不同实施例或示例的特征进行结合和组合。
此外，术语“第一”、“第二”仅用于描述目的，而不能理解为指示或暗示相对重要性或者隐含指明所指示的技术特征的数量。由此，限定有“第一”、“第二”的特征可以明示或者隐含地包括至少一个该特征。在本申请的描述中，“N个”的含义是至少两个，例如两个，三个等，除非另有明确具体的限定。
流程图中或在此以其他方式描述的任何过程或方法描述可以被理解为，表示包括一个或更N个用于实现定制逻辑功能或过程的步骤的可执行指令的代码的模块、片段或部分，并且本申请的优选实施方式的范围包括另外的实现，其中可以不按所示出或讨论的顺序，包括根据所涉及的功能按基本同时的方式或按相反的顺序，来执行功能，这应被本申请的实施例所属技术领域的技术人员所理解。
应当理解，本申请的各部分可以用硬件、软件、固件或它们的组合来实现。在上述实施方式中，N个步骤或方法可以用存储在存储器中且由合适的指令执行系统执行的软件或固件来实现。如，如果用硬件来实现和在另一实施方式中一样，可用本领域公知的下列技术中的任一项或他们的组合来实现：具有用于对数据信号实现逻辑功能的逻辑门电路的离散逻辑电路，具有合适的组合逻辑门电路的专用集成电路，可编程门阵列(PGA)，现场可编程门阵列(FPGA)等。
本技术领域的普通技术人员可以理解实现上述实施例方法携带的全部或部分步骤是可 以通过程序来指令相关的硬件完成，所述的程序可以存储于一种计算机可读存储介质中，该程序在执行时，包括方法实施例的步骤之一或其组合。

Claims (10)

  1. 一种人-车-路微观交通系统建模及风险辨识方法，其特征在于，包括以下步骤：
    利用不同类型的车与道路交通参与者之间的交互作用构建车-路动态交互作用模型，得到车路交互产生的潜在事故后果；
    利用人车系统交互输出的行驶轨迹分布构建人-车动态交互作用模型，得到人车交互产生的行为不确定性；
    利用驾驶人视觉特性对驾驶人驾驶过程中观察道路周边环境的兴趣区域构建人-路动态交互作用模型，得到人路交互过程驾驶人风险敏感度差异；
    利用聚类思想将驾驶人驾驶习性的特征规律和差异性进行表征，得到驾驶人个性化特性；以及
    根据所述车路交互产生的潜在事故后果、所述行为不确定性、所述人路交互过程驾驶人风险敏感度差异及所述驾驶人个性化特性构建人-车-路闭环动力学系统，生成风险辨识结果。
  2. 根据权利要求1所述的方法，其特征在于，所述人-车-路闭环动力学系统表示为：
    Figure PCTCN2022079109-appb-100001
    其中，S DVR为人-车-路闭环动力学系统，γ i(t)为所述人路交互过程驾驶人风险敏感度差异中的有效响应，D i为所述驾驶人个性化特性，δ(t)为车辆转向角度，P x,y*(δ(t))为所述行为不确定性，r′为所述人路交互过程驾驶人风险敏感度差异中的驾驶兴趣区域半径，
    Figure PCTCN2022079109-appb-100002
    为所述车路交互产生的潜在事故后果，ω为转向，ω＝0为车辆直行，ω>0代表左转，ω<0代表右转，Δδ为在某一时间段Δt内转向角的增量，IN为自车i与道路交通参与者j进行交互的交互矩阵，m i,m j和v i,v j分别是自车i与另一个道路交通参与者j的质量和速度，TTC是避撞时间，ξ(v i)＝a/b为，a为椭圆长轴长度，b为椭圆短轴长度，v i(t)为自车i的速度，
    Figure PCTCN2022079109-appb-100003
    为驾驶人的视角函数。
  3. 根据权利要求1所述的方法，其特征在于，所述利用不同类型的车与道路交通参与者之间的交互作用构建车-路动态交互作用模型，得到车路交互产生的潜在事故后果，包括：
    计算场景中自车与不同道路交通参与者的交互数量；
    计算交通环境中两个对象之间的交互后果；
    基于所述交互数量和所述交互后果，生成在多个交通参与者与自车进行交互过程中叠 加交互累计后果得到所述车路交互产生的潜在事故后果
    Figure PCTCN2022079109-appb-100004
    Figure PCTCN2022079109-appb-100005
    其中，IN为自车i与道路交通参与者j进行交互的交互矩阵，m i,m j和v i,v j分别是自车i与另一个道路交通参与者j的质量和速度。
  4. 根据权利要求1所述的方法，其特征在于，所述利用人车系统交互输出的行驶轨迹分布构建人-车动态交互作用模型，得到人车交互产生的行为不确定性，包括：
    根据车辆运动学模型和转弯半径计算等效线性两轮车模型，并根据指令转向角计算预测位置；
    基于采集的实际驾驶实验数据对驾驶人-车辆系统的不确定性运动进行预测；
    基于所述预测位置和所述不确定性运动，得到所述车辆转向角的转角角度的高斯正态分布，并基于所述高斯正态分布进行参数确定，确定所述人车交互产生的行为不确定性P x,y(δ(t))：
    Figure PCTCN2022079109-appb-100006
    其中，δ(t)为车辆转向角度，σ为服从正态分布的数据分布的离散程度，μ为服从正态分布的随机变量的均值。
  5. 根据权利要求1所述的方法，其特征在于，所述利用驾驶人视觉特性对驾驶人驾驶过程中观察道路周边环境的兴趣区域构建人-路动态交互作用模型，得到人路交互过程驾驶人风险敏感度差异，包括：
    利用驾驶人的视觉范围计算驾驶人正常驾驶过程中的有效响应；
    利用驾驶人视觉椭圆形分布特性获取驾驶人动态视觉范围的等势线长度lρ：
    Figure PCTCN2022079109-appb-100007
    其中，r′为兴趣区域大小的半径，a为椭圆长轴长度，b为椭圆短轴长度，a和b取值与自车i车速度v i(t)相关，
    Figure PCTCN2022079109-appb-100008
    为驾驶人的视角函数；
    利用驾驶人视角变化对驾驶人的影响来感知车辆相对速度，得到驾驶人动态感知视野影响，生成驾驶过程中观察道路周边环境的兴趣区域大小的半径：
    Figure PCTCN2022079109-appb-100009
    其中，TTC是避撞时间，ξ(v i)＝a/b，v i(t)为自车i的速度；
    根据所述有效响应以及所述兴趣区域大小的半径确定所述人路交互过程驾驶人风险敏 感度差异。
  6. 根据权利要求1所述的方法，其特征在于，所述利用聚类思想将驾驶人驾驶习性的特征规律和差异性进行表征，得到驾驶人个性化特性，包括：
    区分每类驾驶人在同一场景下的操纵行为和轨迹相似度，基于非监督聚类方法得到驾驶人的个性化表征；
    基于所述驾驶人的个性化表征得到每个驾驶人的特性表征，基于距离聚类中心的距离得到驾驶特性，确定所述驾驶人个性化特性。
  7. 一种人-车-路微观交通系统建模及风险辨识装置，其特征在于，包括：
    第一建模模块，用于利用不同类型的车与道路交通参与者之间的交互作用构建车-路动态交互作用模型，得到车路交互产生的潜在事故后果；
    第二建模模块，用于利用人车系统交互输出的行驶轨迹分布构建人-车动态交互作用模型，得到人车交互产生的行为不确定性；
    第三建模模块，用于利用驾驶人视觉特性对驾驶人驾驶过程中观察道路周边环境的兴趣区域构建人-路动态交互作用模型，得到人路交互过程驾驶人风险敏感度差异；
    表征模块，用于利用聚类思想将驾驶人驾驶习性的特征规律和差异性进行表征，得到驾驶人个性化特性；以及
    辨识模块，用于根据所述车路交互产生的潜在事故后果、所述行为不确定性、所述人路交互过程驾驶人风险敏感度差异及所述驾驶人个性化特性构建人-车-路闭环动力学系统，生成风险辨识结果。
  8. 根据权利要求7所述的装置，其特征在于，所述人-车-路闭环动力学系统表示为：
    Figure PCTCN2022079109-appb-100010
    其中，S DVR为人-车-路闭环动力学系统，γ i(t)为所述人路交互过程驾驶人风险敏感度差异中的有效响应，D i为所述驾驶人个性化特性，δ(t)为车辆转向角度，P x,y*(δ(t))为所述行为不确定性，r′为所述人路交互过程驾驶人风险敏感度差异中的驾驶兴趣区域半径，
    Figure PCTCN2022079109-appb-100011
    为所述车路交互产生的潜在事故后果，ω为转向，ω＝0为车辆直行，ω>0代表左转，ω<0代表右转，Δδ为在某一时间段Δt内转向角的增量，IN为自车i与道路交通参与者j进行交互的交互矩阵，m i,m j和v i,v j分别是自车i与另一个道路交通参与者j的质量和速度，TTC是避撞时间，ξ(v i)＝a/b，a为椭圆长轴长度，b为椭圆短轴长度，v i(t)为自车i的速度，
    Figure PCTCN2022079109-appb-100012
    为驾驶人的视角函数。
  9. 一种电子设备，其特征在于，包括：存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机程序，所述处理器执行所述程序，以实现如权利要求1-6任一项所述的人-车-路微观交通系统建模及风险辨识方法。
  10. 一种计算机可读存储介质，其上存储有计算机程序，其特征在于，该程序被处理器执行，以用于实现如权利要求1-6任一项所述的人-车-路微观交通系统建模及风险辨识方法。
PCT/CN2022/079109 2021-09-26 2022-03-03 人-车-路微观交通系统建模及风险辨识方法及装置 Ceased WO2023045265A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP2024516526A JP7717409B2 (ja) 2021-09-26 2022-03-03 人－車－道路ミクロ交通システムのモデリングとリスク識別方法及び装置

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202111131805.6A CN113946943B (zh) 2021-09-26 2021-09-26 人-车-路微观交通系统建模及风险辨识方法及装置
CN202111131805.6 2021-09-26

Publications (1)

Publication Number Publication Date
WO2023045265A1 true WO2023045265A1 (zh) 2023-03-30

Family

ID=79329498

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/079109 Ceased WO2023045265A1 (zh) 2021-09-26 2022-03-03 人-车-路微观交通系统建模及风险辨识方法及装置

Country Status (3)

Country Link
JP (1) JP7717409B2 (zh)
CN (1) CN113946943B (zh)
WO (1) WO2023045265A1 (zh)

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116386008A (zh) * 2023-04-06 2023-07-04 吉林大学 一种考虑运动不确定性的时空耦合智能汽车风险评估系统
CN116811854A (zh) * 2023-06-29 2023-09-29 北京航空航天大学杭州创新研究院 汽车运行轨迹确定方法、装置、汽车和存储介质
CN117315936A (zh) * 2023-09-27 2023-12-29 杭州电子科技大学 一种街景空间多模态融合的路口场景安全风险量化方法
CN117953444A (zh) * 2024-03-07 2024-04-30 西南交通大学 一种弱势道路使用者事故风险评估方法
CN118411843A (zh) * 2024-04-23 2024-07-30 江苏大学 基于数据机理耦合建模的车路协同控制架构系统及构建方法
CN118942277A (zh) * 2024-07-03 2024-11-12 淮阴工学院 一种斑马线人车碰撞预警方法及系统
CN119152688A (zh) * 2024-11-11 2024-12-17 浙江大华系统工程有限公司 一种危险个体预测方法、装置及电子设备
CN119312673A (zh) * 2024-09-27 2025-01-14 同济大学 一种基于时空图Transformer的交织区微观仿真系统
CN120145638A (zh) * 2025-02-11 2025-06-13 清华大学 4d风险占用数据集的生成方法、装置、设备及介质

Families Citing this family (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113946943B (zh) * 2021-09-26 2022-06-21 清华大学 人-车-路微观交通系统建模及风险辨识方法及装置
CN115440088B (zh) * 2022-07-25 2024-05-14 重庆大学 一种基于场的智能车辆换道决策方法
CN115630200B (zh) * 2022-11-01 2025-03-14 长安大学 一种隧道路段驾驶行为谱构建方法及系统
CN116749925B (zh) * 2023-06-20 2024-05-07 哈尔滨工业大学 面向人机混驾的仿人工驾驶证据动态累积制动方法及系统
CN116968730B (zh) * 2023-06-25 2024-03-19 清华大学 高风险场景下驾驶人风险响应与主动决策方法及装置
CN116776204B (zh) * 2023-06-26 2024-02-23 清华大学 驾驶人风险敏感度差异化表征方法、装置、设备及介质
CN116578103B (zh) * 2023-07-13 2023-10-03 清华大学 极限工况下人-车-路闭环系统运行风险评测方法
CN120125026B (zh) * 2025-02-25 2025-11-18 机械工业仪器仪表综合技术经济研究所 基于多元向量积的数字化车间风险评估方法及系统

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108648447A (zh) * 2018-05-08 2018-10-12 清华大学 基于行车安全场的车辆智能安全决策方法
US20190100216A1 (en) * 2017-09-29 2019-04-04 Denso International America, Inc. Risk Assessment System For Assessing Current Driver Behavior Relative to Past Behavior and Behaviors of Other Drivers
CN109740840A (zh) * 2018-11-23 2019-05-10 深圳市大东车慧科技股份有限公司 一种基于人、车、路和环境的多元分析驾驶风险评估系统
CN111523822A (zh) * 2020-05-06 2020-08-11 重庆文理学院 一种综合考虑人-车-路各因素的行车风险统一量化方法
CN113946943A (zh) * 2021-09-26 2022-01-18 清华大学 人-车-路微观交通系统建模及风险辨识方法及装置

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP3882797B2 (ja) * 2003-08-08 2007-02-21 日産自動車株式会社 車両用運転操作補助装置および車両用運転操作補助装置を備える車両
US11364899B2 (en) * 2017-06-02 2022-06-21 Toyota Motor Europe Driving assistance method and system
CN108639059B (zh) 2018-05-08 2019-02-19 清华大学 基于最小作用量原理的驾驶人操控行为量化方法及装置
CN111126868B (zh) * 2019-12-30 2023-07-04 中南大学 一种道路交通事故发生风险确定方法及系统
CN112677983B (zh) * 2021-01-07 2022-04-12 浙江大学 一种识别驾驶员驾驶风格的系统

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190100216A1 (en) * 2017-09-29 2019-04-04 Denso International America, Inc. Risk Assessment System For Assessing Current Driver Behavior Relative to Past Behavior and Behaviors of Other Drivers
CN108648447A (zh) * 2018-05-08 2018-10-12 清华大学 基于行车安全场的车辆智能安全决策方法
CN109740840A (zh) * 2018-11-23 2019-05-10 深圳市大东车慧科技股份有限公司 一种基于人、车、路和环境的多元分析驾驶风险评估系统
CN111523822A (zh) * 2020-05-06 2020-08-11 重庆文理学院 一种综合考虑人-车-路各因素的行车风险统一量化方法
CN113946943A (zh) * 2021-09-26 2022-01-18 清华大学 人-车-路微观交通系统建模及风险辨识方法及装置

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
HUANG, HEYE ET AL.: "A probabilistic risk assessment framework considering lane-changing behavior interaction", SCIENCE CHINA(INFORMATION SCIENCES), vol. 63, no. 09, 17 August 2020 (2020-08-17), XP037224095, ISSN: 1674-733X, DOI: 10.1007/s11432-019-2983-0 *

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116386008A (zh) * 2023-04-06 2023-07-04 吉林大学 一种考虑运动不确定性的时空耦合智能汽车风险评估系统
CN116811854A (zh) * 2023-06-29 2023-09-29 北京航空航天大学杭州创新研究院 汽车运行轨迹确定方法、装置、汽车和存储介质
CN117315936A (zh) * 2023-09-27 2023-12-29 杭州电子科技大学 一种街景空间多模态融合的路口场景安全风险量化方法
CN117953444A (zh) * 2024-03-07 2024-04-30 西南交通大学 一种弱势道路使用者事故风险评估方法
CN118411843A (zh) * 2024-04-23 2024-07-30 江苏大学 基于数据机理耦合建模的车路协同控制架构系统及构建方法
CN118942277A (zh) * 2024-07-03 2024-11-12 淮阴工学院 一种斑马线人车碰撞预警方法及系统
CN119312673A (zh) * 2024-09-27 2025-01-14 同济大学 一种基于时空图Transformer的交织区微观仿真系统
CN119152688A (zh) * 2024-11-11 2024-12-17 浙江大华系统工程有限公司 一种危险个体预测方法、装置及电子设备
CN120145638A (zh) * 2025-02-11 2025-06-13 清华大学 4d风险占用数据集的生成方法、装置、设备及介质

Also Published As

Publication number Publication date
CN113946943B (zh) 2022-06-21
CN113946943A (zh) 2022-01-18
JP2024533514A (ja) 2024-09-12
JP7717409B2 (ja) 2025-08-04

Similar Documents

Publication Publication Date Title
CN113946943B (zh) 人-车-路微观交通系统建模及风险辨识方法及装置
Tawari et al. Looking-in and looking-out vision for urban intelligent assistance: Estimation of driver attentive state and dynamic surround for safe merging and braking
EP1332910B1 (en) Method and system for vehicle operator assistance improvement
CN113085873B (zh) 驾驶策略的获取方法、装置、计算机设备和存储介质
CN114117829B (zh) 极限工况下人-车-路闭环系统动力学建模方法及系统
Kim et al. Trajectory planning and control of autonomous vehicles for static vehicle avoidance in dynamic traffic environments
GB2601841A (en) Scenario-based behavior specification and validation
CN105966396A (zh) 一种基于驾驶员避撞行为的车辆避撞控制方法
Zheng et al. Intelligent decision-making method for vehicles in emergency conditions based on artificial potential fields and finite state machines
Wang et al. Model predictive control–based cooperative lane change strategy for improving traffic flow
CN113370980B (zh) 极限工况下车辆失稳与碰撞综合风险评估方法及装置
CN117116072A (zh) 一种高速公路协同决策安全行车方法及系统
CN110660271A (zh) 在更大距离上识别对于行驶重要相关的情况
Bellet et al. A computational model of the car driver interfaced with a simulation platform for future Virtual Human Centred Design applications: COSMO-SIVIC
CN110646007A (zh) 一种基于形式化表示的车辆驾驶方法
Kankoriya et al. Advance vehicle advanced driver assistance systems: working & features ADAS a path towards intelligent vehicles
US20200211388A1 (en) Augmented reality based driver guidance system
Hunter et al. Cooperative Vehicle–Highway Automation (CVHA) Technology: Simulation of Benefits and Operational Issues
Kumar et al. Advancement of advanced driver assistance system in automobiles through IoT implementation and integration
Zheng et al. A novel road traffic risk modeling approach based on the traffic safety field concept
Deng et al. Analysis of vehicle assisted lane change system and autonomous lane change model
Peng et al. Hierarchical motion planning system with consideration of the dynamic lane-changing behaviour
TWI887807B (zh) 車輛行駛風險檢測方法及其系統
Yuan et al. Optimization of forward collision warning algorithm and driver-in-the-loop validation for intelligent vehicles under low-visibility conditions
Magán et al. Fuzzy Alarm System based on Human-centered Approach.

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22871330

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2024516526

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 22871330

Country of ref document: EP

Kind code of ref document: A1