WO2023141940A1 - 一种智能驾驶方法、装置及车辆 - Google Patents

一种智能驾驶方法、装置及车辆 Download PDF

Info

Publication number
WO2023141940A1
WO2023141940A1 PCT/CN2022/074581 CN2022074581W WO2023141940A1 WO 2023141940 A1 WO2023141940 A1 WO 2023141940A1 CN 2022074581 W CN2022074581 W CN 2022074581W WO 2023141940 A1 WO2023141940 A1 WO 2023141940A1
Authority
WO
WIPO (PCT)
Prior art keywords
target vehicle
vehicle
information
degree
influence
Prior art date
Application number
PCT/CN2022/074581
Other languages
English (en)
French (fr)
Inventor
吴晗
Original Assignee
华为技术有限公司
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 华为技术有限公司 filed Critical 华为技术有限公司
Priority to PCT/CN2022/074581 priority Critical patent/WO2023141940A1/zh
Publication of WO2023141940A1 publication Critical patent/WO2023141940A1/zh

Links

Images

Classifications

    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W30/00—Purposes of road vehicle drive control systems not related to the control of a particular sub-unit, e.g. of systems using conjoint control of vehicle sub-units, or advanced driver assistance systems for ensuring comfort, stability and safety or drive control systems for propelling or retarding the vehicle
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W50/00—Details of control systems for road vehicle drive control not related to the control of a particular sub-unit, e.g. process diagnostic or vehicle driver interfaces
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
    • G05D1/02—Control of position or course in two dimensions
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled

Definitions

  • the present application relates to the technical field of intelligent driving, in particular to an intelligent driving method, device and vehicle.
  • CIPV Cert In-Path Vehicle
  • the technical problem to be solved by the embodiments of the present application is to provide an intelligent driving method, device and vehicle to smoothly complete the longitudinal control of the own vehicle and improve driving safety and user comfort.
  • the embodiment of the present application provides a smart driving method, which may include:
  • a risk weight coefficient of each target vehicle is determined according to the second driving state information of each target vehicle and the first driving state information, and the risk weight coefficient is used to adjust the driving state of the own vehicle.
  • the relationship between the multiple target vehicles and the self-vehicle is obtained
  • the risk weight coefficient so that the driving state of the self-vehicle can be controlled according to the risk weight coefficient of multiple target vehicles. Due to the comprehensive consideration of the information of multiple target vehicles within the preset range, driving safety is improved and avoid The sudden acceleration and deceleration caused by a single target vehicle in the process of switching and releasing are eliminated, thus improving the user's comfort.
  • the determining the risk weight coefficient of each target vehicle according to the second driving state information of each target vehicle and the first driving state information includes:
  • the degree of influence satisfies a preset corresponding relationship with the second driving state information and the first driving state information of the target vehicle.
  • Evaluating the risk weight coefficient of each target vehicle through the degree of influence of each target vehicle on the self-vehicle can improve the comprehensiveness and accuracy of the comprehensive control.
  • the first and/or second driving state information includes road condition information
  • the road condition information includes at least one of the following:
  • the location information of the target vehicle can be determined through the curb information and lane line information in the road condition information, and then the determination of the risk weight coefficient can be completed more accurately. And comprehensive consideration of road condition information can improve the safety of self-vehicle control.
  • the road condition information is used to characterize the driving intention of each target vehicle.
  • the method further includes:
  • the driving intention of the first target vehicle is determined according to the direction in which the degree of comprehensive influence decreases.
  • the driving intention of the first target vehicle affects the risk weight coefficient in the following manner:
  • the risk weight coefficient of the first target vehicle is increased
  • the risk weight coefficient of the first target vehicle is reduced.
  • the method further includes:
  • the comprehensive acceleration of the own vehicle is calculated according to the expected acceleration of each target vehicle and the risk weight coefficient of each target vehicle.
  • the controller can control the driving state of the vehicle according to the integrated acceleration. Since the expected acceleration and risk weight coefficients of multiple target vehicles are considered in the comprehensive acceleration, more accurate and safe control of the ego vehicle can be realized.
  • the method further includes:
  • the information of the single virtual target vehicle is determined according to the risk weight coefficient and the second driving state information of each target vehicle;
  • the virtual driving scene includes a planned driving trajectory, and the planned driving trajectory is obtained according to the first driving state information of the own vehicle and the second driving state information of each target vehicle.
  • the first driving state information includes the speed information of the own vehicle and acceleration information
  • the second driving state information includes the target vehicle's position information (xi , y i ), speed information and acceleration information
  • x′ i , y′ i are calculated according to the relative velocity and acceleration in the transverse direction and longitudinal direction as follows:
  • Parameters ⁇ x , ⁇ y , ⁇ x , ⁇ y can respectively adjust the degree of influence of lateral velocity, longitudinal velocity, lateral acceleration and longitudinal acceleration on normalization coefficient ⁇ x , ⁇ y , the larger ⁇ x , ⁇ y the smaller the influence range of potential energy .
  • the separately calculating the degree of influence of each piece of information contained in the road condition information on the first target vehicle includes:
  • the degree of influence of the roadside line or lane line on the first target vehicle satisfies the following formula:
  • the coefficient ⁇ lane is an adjustment coefficient for adjusting the degree of influence of the lane line/road; ⁇ lane is used for adjusting the size of the attenuation of the degree of influence; y lane is the lateral distance between the first target vehicle and the lane line/road;
  • the degree of influence of the passable obstacle on the first target vehicle satisfies the following formula:
  • the coefficient ⁇ obs is the adjustment coefficient for adjusting the influence degree of the passable obstacle
  • ⁇ x , ⁇ y are the adjustment coefficients respectively affecting the attenuation speed of the longitudinal and lateral influence degree
  • x obs , y obs are the passable obstacles object location.
  • the estimating the degree of willingness of the first target vehicle to drive forward satisfies the following formula:
  • ⁇ attr is an adjustment coefficient for adjusting the degree of willingness.
  • the gradient descent method is used to search for the direction of reducing the comprehensive influence degree, which satisfies the following formula:
  • ⁇ is the gradient descent step size.
  • the calculation of the expected acceleration of each target vehicle according to the car-following model satisfies the following formula:
  • v is the speed of the self-vehicle
  • ⁇ v is the speed difference between the self-vehicle and the vehicle in front of the current lane
  • v des is the expected speed
  • ⁇ is the calibration parameter
  • a max is the maximum acceleration
  • s * is the expected distance:
  • s 0 is the minimum distance of following and stopping; T is the following distance; b is the comfortable deceleration.
  • the comprehensive acceleration of the self-vehicle is calculated according to the risk weight coefficient, which satisfies the following formula:
  • a multi ⁇ i ⁇ a i ,
  • i is the serial number of the target vehicle
  • ⁇ i is the risk weight coefficient of the target vehicle i
  • a i is the expected acceleration of the target vehicle i.
  • the information of the virtual target vehicle includes position information, speed information and acceleration information of the virtual target vehicle, and the position information, speed information and acceleration information of the virtual target vehicle satisfy the following formula :
  • i is the serial number of the target vehicle
  • ⁇ i is the risk weight coefficient of the target vehicle i
  • (x virtual , y virtual ) is the position information of the virtual target vehicle
  • the calibration of various adjustment coefficients is determined according to external input parameters.
  • the calibration of various adjustment coefficients is determined according to the following objective function:
  • the objective function is used to determine the descending direction of the minimized influence on the actual driving direction of the ego vehicle at time t, wherein the direction vector of the ego vehicle at time t is Gradient descent direction of the degree of influence on the self-vehicle
  • an intelligent driving device which may include:
  • an acquisition unit configured to acquire the first driving state information of the own vehicle and the second driving state information of a plurality of target vehicles, the target vehicles being located within a preset range;
  • a processing unit configured to determine a risk weight coefficient of each target vehicle according to the second driving state information of each target vehicle and the first driving state information, and the risk weight coefficient is used to adjust the driving state of the own vehicle.
  • the processing unit is specifically configured to:
  • the degree of influence satisfies a preset corresponding relationship with the second driving state information and the first driving state information of the target vehicle.
  • the first and/or second driving state information includes road condition information
  • the road condition information includes at least one of the following:
  • the road condition information is used to characterize the driving intention of each target vehicle.
  • the processing unit when determining the driving intention of the first target vehicle among the target vehicles according to the road condition information, is configured to:
  • the driving intention of the first target vehicle is determined according to the direction in which the degree of comprehensive influence decreases.
  • the driving intention of the first target vehicle affects the risk weight coefficient in the following manner:
  • the risk weight coefficient of the first target vehicle is increased
  • the risk weight coefficient of the first target vehicle is reduced.
  • the processing unit is further configured to:
  • the comprehensive acceleration of the own vehicle is calculated according to the expected acceleration of each target vehicle and the risk weight coefficient of each target vehicle.
  • the processing unit is further configured to:
  • the information of the single virtual target vehicle is determined according to the risk weight coefficient and the second driving state information of each target vehicle;
  • the virtual driving scene includes a planned driving trajectory, and the planned driving trajectory is obtained according to the first driving state information of the own vehicle and the second driving state information of each target vehicle.
  • the first driving state information includes the speed information of the own vehicle and acceleration information
  • the second driving state information includes the target vehicle's position information (xi , y i ), speed information and acceleration information
  • x′ i , y′ i are calculated according to the relative velocity and acceleration in the transverse direction and longitudinal direction as follows:
  • Parameters ⁇ x , ⁇ y , ⁇ x , ⁇ y can respectively adjust the degree of influence of lateral velocity, longitudinal velocity, lateral acceleration and longitudinal acceleration on normalization coefficient ⁇ x , ⁇ y , the larger ⁇ x , ⁇ y the smaller the influence range of potential energy .
  • the processing unit calculates the degree of influence of the roadside line or lane line on the first target vehicle, satisfying The following formula:
  • the coefficient ⁇ lane is an adjustment coefficient for adjusting the degree of influence of the lane line/road; ⁇ lane is used for adjusting the size of the attenuation of the degree of influence; y lane is the lateral distance between the first target vehicle and the lane line/road;
  • the degree of influence of the passable obstacle on the first target vehicle satisfies the following formula:
  • the coefficient ⁇ obs is the adjustment coefficient for adjusting the influence degree of the passable obstacle
  • ⁇ x , ⁇ y are the adjustment coefficients respectively affecting the attenuation speed of the longitudinal and lateral influence degree
  • x obs , y obs are the passable obstacles object location.
  • the processing unit estimates the degree of willingness of the first target vehicle to drive forward, which satisfies the following formula:
  • ⁇ attr is an adjustment coefficient for adjusting the degree of willingness.
  • the processing unit calculates the The comprehensive impact degree of the first target vehicle satisfies the following formula:
  • the processing unit searches for a direction in which the degree of comprehensive influence is reduced by using a gradient descent method, which satisfies the following formula:
  • ⁇ is the gradient descent step size.
  • the processing unit respectively calculates the expected acceleration of each target vehicle according to the car-following model, which satisfies the following formula:
  • v is the speed of the self-vehicle
  • ⁇ v is the speed difference between the self-vehicle and the vehicle in front of the current lane
  • v des is the expected speed
  • ⁇ is the calibration parameter
  • a max is the maximum acceleration
  • s * is the expected distance:
  • s 0 is the minimum distance of following and stopping; T is the following distance; b is the comfortable deceleration.
  • the processing unit calculates the comprehensive acceleration of the own vehicle according to the expected acceleration of each target vehicle and the risk weight coefficient of each target vehicle, which satisfies the following formula:
  • a multi ⁇ i ⁇ a i ,
  • i is the serial number of the target vehicle
  • ⁇ i is the risk weight coefficient of the target vehicle i
  • a i is the expected acceleration of the target vehicle i.
  • the information of the virtual target vehicle includes position information, speed information and acceleration information of the virtual target vehicle, and the position information, speed information and acceleration information of the virtual target vehicle satisfy the following formula :
  • i is the serial number of the target vehicle
  • ⁇ i is the risk weight coefficient of the target vehicle i
  • (x virtual , y virtual ) is the position information of the virtual target vehicle
  • the calibration of various adjustment coefficients is determined according to external input parameters.
  • the calibration of various adjustment coefficients is determined according to the following objective function:
  • the objective function is used to determine the descending direction of the minimized influence on the actual driving direction of the ego vehicle at time t, wherein the direction vector of the ego vehicle at time t is Gradient descent direction of the degree of influence on the self-vehicle
  • an intelligent driving device which may include:
  • the embodiments of the present application provide a computer-readable storage medium, the computer-readable storage medium stores instructions, and when it is run on a computer, it can realize the above-mentioned first aspect or any one of the first aspect. Implement the method described in the manner.
  • a computer program product comprising: computer program code, when the computer program code is run on a computer, the computer is made to execute the first aspect or any implementation manner of the first aspect the method described.
  • the embodiment of the present application provides an intelligent driving device, which may be a chip or a system on a chip in a vehicle, and the device includes a processor, the processor is coupled with a memory, and the memory is used to store computer programs or Instructions, the processor is used to execute the computer programs or instructions in the memory, so that the device executes the method described in the first aspect or any implementation manner of the first aspect.
  • the device further includes the memory.
  • the embodiments of the present application provide a vehicle, which may include:
  • FIG. 1 is a schematic structural diagram of a vehicle control system provided by an embodiment of the present application
  • FIG. 2 is a schematic flow diagram of an intelligent driving method provided in an embodiment of the present application.
  • FIG. 3 is a schematic flow diagram of another intelligent driving method provided in the embodiment of the present application.
  • FIG. 4 is a schematic flow diagram of another intelligent driving method provided in the embodiment of the present application.
  • FIG. 5 is a schematic flow diagram of another intelligent driving method provided in the embodiment of the present application.
  • Fig. 6 is a schematic diagram of the comparison of the simulation results of the intelligent driving method and the single-target longitudinal control provided by the embodiment of the present application;
  • FIG. 7 is a schematic diagram of the composition of an intelligent driving device provided by an embodiment of the present application.
  • FIG. 8 is a schematic composition diagram of another intelligent driving device provided by an embodiment of the present application.
  • references in this application to "including” and “having” and any variations thereof are intended to cover a non-exclusive inclusion.
  • a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but optionally also includes steps or units that are not listed, or optionally Other steps or elements inherent to these processes, methods, products or devices are also included.
  • Level L0 driving automation (emergency assistance): The driving automation system cannot continuously perform vehicle lateral or longitudinal motion control in dynamic driving tasks, but has the ability to continuously perform some target and event detection and response in dynamic driving tasks.
  • Level L1 driving automation (partial driving assistance): The driving automation system continuously performs the vehicle lateral or longitudinal motion control in dynamic driving tasks within its design operating conditions, and has a part adapted to the executed vehicle lateral or longitudinal motion control Target and incident detection and response capabilities.
  • Level 2 driving automation combined driving assistance: The driving automation system continuously performs the vehicle lateral and longitudinal motion control in dynamic driving tasks within its design operating conditions, and has a part adapted to the executed vehicle lateral and longitudinal motion control Target and incident detection and response capabilities.
  • Level 3 driving automation (conditional autonomous driving): The driving automation system continuously performs all dynamic driving tasks within its design operating conditions.
  • Level 4 driving automation (highly automated driving): The driving automation system continuously performs all dynamic driving tasks and performs dynamic driving task takeover within its design operating conditions.
  • Level 5 driving automation full autonomous driving: The driving automation system continuously performs all dynamic driving tasks and performs dynamic driving task takeover under any drivable conditions.
  • the intelligent driving method in the embodiment of the present application can be applied to domestic L1/L2 level automatic driving, and can also be applied to domestic L2 level or above automatic driving scenarios. And it is also applicable to the L1-L5 level of automatic driving formulated by the Society of Automotive Engineers (SAE).
  • SAE Society of Automotive Engineers
  • FIG. 1 is a schematic structural diagram of a vehicle control system provided by an embodiment of the present application. May include:
  • An intelligent driving device 10 a radar 20 , a camera 30 , an inertial measurement unit (Inertial Measurement Unit, IMU) 40 and a controller 50 .
  • IMU Inertial Measurement Unit
  • the intelligent driving device 10 is used to perceive the situation of the vehicle and the surrounding environment through various sensors carried by the vehicle.
  • the driving state information of the vehicle such as speed information and acceleration information can be obtained through the IMU;
  • the radar 20 perceives vehicles and obstacles in the surrounding environment;
  • the camera 30 can also capture road condition information such as surrounding vehicle information, lane line information, and roadside information.
  • the intelligent driving device 10 can finally determine the driving state of the own vehicle according to the obtained second driving state information (including information such as position, speed and acceleration) of multiple target vehicles within the preset range.
  • the type of radar 20 may be ultrasonic radar, microwave radar or laser radar, etc.
  • the number of radars mounted on the self-vehicle may be one or more. It can be used to detect obstacles, predict collisions, adaptive cruise control, etc.
  • the radar 20 can be installed on the central part and the corner of the front surface of the self-vehicle, the central part and the corner of the rear surface, etc., and can emit electromagnetic waves or lasers within a predetermined angle range in the front area of the radar 20, and can receive signals from The echoes reflected by surrounding objects near the car can detect the angle, distance, speed, acceleration, etc. between the car and various surrounding objects, and send this information to the intelligent driving device 10 .
  • the type of the camera 30 may be an infrared camera, a complementary metal oxide semiconductor (Complementary Metal Oxide Semiconductor, CMOS) camera, a charge coupled device (charge coupled device, CCD) camera, or a laser camera.
  • CMOS Complementary Metal Oxide Semiconductor
  • CCD charge coupled device
  • the number of cameras mounted on the vehicle can be one or more. It can be used to take pictures of the environment and vehicles around the car.
  • the camera 30 can also be mounted on the center and corners of the front surface of the own vehicle, the center and corners of the rear surface, etc., and can also be installed on the upper end of the windshield of the host vehicle.
  • Various types of light are sensed and projected within a predetermined angle and predetermined distance range in the front area, and images of external objects around the vehicle can be acquired, and the acquired images can be sent to the smart driving device 10.
  • the intelligent driving device 10 may include an acquisition unit 11 and a processing unit 12, the acquisition unit 11 is used to acquire data from various sensors, and the processor is used to process the acquired data.
  • the intelligent driving device 10 may also include a memory (eg, DB) such as a read-only memory (ROM) or a random access memory (RAM), which can store acquired data and program codes that the processing unit 12 needs to execute. New data generated after the processing unit 12 processes the acquired data may be stored.
  • DB such as a read-only memory (ROM) or a random access memory (RAM)
  • the acquiring unit 11 can acquire the first driving state information of the own vehicle through the IMU40, and can also acquire the second driving state information of the target vehicle within a preset range through sensors such as the radar 20 and the camera 30;
  • the processing unit 12 can determine the degree of influence of each target vehicle on the own vehicle according to the information acquired by the acquisition unit 11, and determine the risk weight coefficient of each target vehicle relative to the own vehicle, thereby outputting smoother and more accurate control information such as integrated acceleration or The travel trajectory and the like are given to the controller 50 .
  • the controller 50 may be a proportional-integral-derivative (Proportion Integration Differentiation, PID) controller or other types of controllers.
  • the controller 50 may include a memory (for example, DB) such as a read only memory (ROM) or a random access memory (RAM), may store various control data and control program codes, and may further include a processor such as a CPU , so that the controller 50 can execute various control programs.
  • DB memory
  • ROM read only memory
  • RAM random access memory
  • CPU random access memory
  • Receive the control information output by the intelligent driving device 10 and adjust the driving state of the own vehicle according to the control information. For example, control the speed and/or acceleration of the own vehicle according to the comprehensive acceleration or driving trajectory output by the intelligent driving device.
  • any sensor that can be used to improve the sensing capability of the self-vehicle such as a light detection sensor and a ranging sensor, can also be mounted on the self-vehicle.
  • any sensor that can be used to improve the sensing capability of the self-vehicle such as a light detection sensor and a ranging sensor, can also be mounted on the self-vehicle.
  • the embodiment of this application does not make any limitation.
  • Fig. 2 is a schematic flow chart of an intelligent driving method provided by the embodiment of the present application; it may include the following steps:
  • the own vehicle can obtain the first driving state information of the own vehicle such as speed information and acceleration information through the IMU; 2. Driving state information such as position information, speed information, acceleration information and road condition information.
  • the target vehicle is located within a preset range and the number is greater than one.
  • the preset range can be set by the controller of the self-vehicle according to externally input data, or can be set by a remote server, and can also be set or adjusted by the user according to his own needs. For example, in rainy and foggy weather where the line of sight is obstructed or when the current user is very concerned about safety, the preset range can be set to be larger, while in the case of normal sightlines or the current user wants to improve the efficiency of the vehicle's data processing, the preset range can be set to The preset range settings are smaller.
  • the first driving state information of the own vehicle may include speed information and acceleration information of the own vehicle, or may also include road condition information; the second driving state information correspondingly exists in multiples according to the number of target vehicles.
  • the second driving state information of each target vehicle may include position information, speed information and acceleration information of the target vehicle, or may also include road condition information.
  • the road condition information includes at least one of the following:
  • the location information may be determined according to roadside information and/or lane line information obtained by the self-vehicle.
  • S202 Determine the risk weight coefficient of each target vehicle according to the second driving state information of each target vehicle and the first driving state information.
  • the risk weight coefficient is used to adjust the driving state of the own vehicle.
  • each target vehicle within the preset range will have different degrees of influence on the own vehicle, for example, whether the target vehicle in front of the current lane of the own vehicle accelerates or decelerates, whether it changes lanes, etc. will affect the own vehicle; Whether the lane on the side or the right side changes lanes will affect the own vehicle, etc.
  • the risk weight coefficient of each target vehicle can be determined according to the influence degree of each target vehicle on the own vehicle
  • the degree of influence satisfies a preset corresponding relationship with the second driving state information and the first driving state information of the target vehicle.
  • the degree of influence of each target vehicle on the self-vehicle satisfies the following formula:
  • x′ i , y′ i are calculated according to the relative velocity and acceleration in the transverse direction and longitudinal direction as follows:
  • Parameters ⁇ x , ⁇ y , ⁇ x , ⁇ y can respectively adjust the degree of influence of lateral velocity, longitudinal velocity, lateral acceleration and longitudinal acceleration on normalization coefficient ⁇ x , ⁇ y , the larger ⁇ x , ⁇ y the smaller the influence range of potential energy .
  • the calculated data of the degree of influence on the self-vehicle are 80, 60, 40, and 20 respectively.
  • the risk weight coefficients obtained after normalization processing are 40%, 30%, 20%, and 10%, respectively.
  • the present application by obtaining the first driving state information of the self-vehicle and the second driving state information of multiple target vehicles within the preset range, and evaluating the multiple target vehicles within the preset range, multiple The risk weight coefficient of a target vehicle to the self-vehicle, so that the driving state of the self-vehicle can be controlled according to the risk weight coefficients of multiple target vehicles. Due to the comprehensive consideration of the information of multiple target vehicles within the preset range, it is improved. Driving safety, and avoiding the sudden acceleration and deceleration caused by a single target vehicle in the process of switching and releasing, thus improving the user's comfort.
  • the road condition information can be used to characterize the driving intention of each target vehicle.
  • the risk weight coefficient can be further adjusted according to the driving intention of the target vehicle.
  • Fig. 3 is a schematic flow diagram of another intelligent driving method provided by the embodiment of the present application; wherein, steps S301-step S302 are the same as step S201-step S202 in Fig. 2, and will not be repeated here. After step S302, the following steps are also included:
  • the degree of influence of other vehicles on the first target vehicle may be calculated by referring to the method of calculating the degree of influence of the target vehicle on the own vehicle in step S202.
  • the calculation may be performed by considering the target vehicle in front of the own vehicle.
  • other vehicles in front of the first target vehicle and other vehicles behind the first target vehicle may be considered for calculation.
  • the degree of influence of the roadside line or lane line on the first target vehicle satisfies the following formula:
  • the coefficient ⁇ lane is an adjustment coefficient for adjusting the degree of influence of the lane line/road; ⁇ lane is used for adjusting the size of the attenuation of the degree of influence; y lane is the lateral distance between the first target vehicle and the lane line/road;
  • the degree of influence of the passable obstacle on the first target vehicle satisfies the following formula:
  • the coefficient ⁇ obs is the adjustment coefficient for adjusting the influence degree of the passable obstacle
  • ⁇ x , ⁇ y are the adjustment coefficients respectively affecting the attenuation speed of the longitudinal and lateral influence degree
  • x obs , y obs are the passable obstacles object location.
  • the degree of willingness is used to characterize the inertia of the first target vehicle traveling forward. For example, if the speed and acceleration of the first target vehicle have not changed or have not changed significantly for a period of time, it can be determined that the first target vehicle has a strong willingness to drive forward; if the lateral acceleration of the first target vehicle occurs If there is a change, there may be the possibility of changing lanes or turning or turning around, and the willingness to drive forward is relatively low.
  • the degree of willingness of the first target vehicle to move forward is estimated to satisfy the following formula:
  • ⁇ attr is an adjustment coefficient for adjusting the degree of willingness.
  • the comprehensive impact degree of the first target vehicle is calculated to satisfy the following formula:
  • the gradient descent method (gradient descent) is often used in machine learning and artificial intelligence to recursively approximate the minimum deviation model. Its calculation process includes finding the minimum value of the objective function or converging to the minimum value through iteration along the direction of gradient descent.
  • the gradient descent method is used to search for the direction in which the degree of comprehensive influence is reduced, which satisfies the following formula:
  • ⁇ is the gradient descent step size.
  • S307. Determine the driving intention of the first target vehicle according to the direction in which the degree of comprehensive influence decreases.
  • the possible future driving direction of the first target vehicle can be estimated by obtaining the direction with reduced comprehensiveness through the gradient descent method. Further, the driving intention of the first target vehicle is determined.
  • the direction in which the degree of comprehensive influence decreases corresponds to the fact that the first target vehicle may drive to the right, and the driving intention of the first target vehicle is to change lanes and enter the lane where the own vehicle is located, then the influence of the first target vehicle on the own vehicle will change at this time big. If the direction in which the degree of comprehensive influence decreases corresponds to the possibility of the first target vehicle driving to the left, and the driving intention of the first target vehicle is to change lanes and leave the lane where the own vehicle is located, then the impact of the first target vehicle on the own vehicle will be reduced at this time .
  • the direction in which the degree of comprehensive influence decreases corresponds to the fact that the first target vehicle may keep moving forward, and the driving intention of the first target vehicle is to maintain the current driving state and drive in the original lane, then the influence of the first target vehicle on the ego vehicle will remain unchanged at this time.
  • the risk weight coefficient of the first target vehicle is increased if the driving intention of the first target vehicle is to drive into the lane where the ego vehicle is currently located.
  • the risk weight coefficient of the first target vehicle is reduced.
  • the risk weight coefficient of the target vehicle within the preset range is normalized again.
  • the calibration of various adjustment coefficients such as ⁇ lane , ⁇ car , ⁇ obs , ⁇ attr , ⁇ lane, ⁇ x , ⁇ y in the embodiment of the present application is determined according to external input parameters.
  • the calibration of various adjustment coefficients in the embodiment of the present application is determined according to the following objective function:
  • the objective function is used to determine the descending direction of the minimized influence on the actual driving direction of the ego vehicle at time t, wherein the direction vector of the ego vehicle at time t is Gradient descent direction of the degree of influence on the self-vehicle
  • the risk weight coefficient of the target vehicle can be further finely adjusted to obtain information that is more in line with future comprehensive road conditions, further improving driving safety Prospect and control.
  • the embodiment of the present application can output the integrated acceleration to the controller so that the controller can control the driving state of the own vehicle.
  • FIG. 4 is a schematic flow diagram of another intelligent driving method provided by the embodiment of the present application; steps S401-step S408 are the same as steps S301-step S308 in FIG. Afterwards, the following steps are also included:
  • Car Following (CF) behavior is the most basic microscopic driving behavior, which describes the interaction between two adjacent vehicles in a platoon driving on a one-way street where overtaking is restricted.
  • Car-following model is to use the method of dynamics to study the corresponding behavior of the following vehicle (Following Vehicle, FV) caused by the change of the motion state of the leading vehicle (Leading Vehicle, LV). Traffic flow characteristics, so as to build a bridge between the driver's microscopic behavior and the traffic macroscopic phenomenon.
  • FV Flullowing Vehicle
  • LV Leading Vehicle
  • Traffic flow characteristics so as to build a bridge between the driver's microscopic behavior and the traffic macroscopic phenomenon.
  • the method in the embodiment of the present application does not limit the specific form or version of the car-following model, and is applicable to various versions of the car-following model that currently exist or will be more advanced in the future.
  • the expected acceleration of each target vehicle is calculated separately according to the car-following model, which satisfies the following formula:
  • v is the speed of the self-vehicle
  • ⁇ v is the speed difference between the self-vehicle and the vehicle in front of the current lane
  • v des is the expected speed
  • ⁇ is the calibration parameter
  • a max is the maximum acceleration
  • s * is the expected distance:
  • s 0 is the minimum distance of following and stopping; T is the following distance; b is the comfortable deceleration.
  • the calculation of integrated acceleration satisfies the following formula:
  • a multi ⁇ i ⁇ a i ,
  • i is the serial number of the target vehicle
  • ⁇ i is the risk weight coefficient of the target vehicle i
  • a i is the expected acceleration of the target vehicle i.
  • the controller can control the driving state of the vehicle according to the integrated acceleration. Since the expected acceleration and risk weight coefficients of multiple target vehicles are considered in the comprehensive acceleration, more accurate and safe control of the ego vehicle can be realized.
  • Fig. 5 is a schematic flow diagram of another intelligent driving method provided by the embodiment of the present application; Step S501-Step S508 is the same as Step S301-Step S308 in Fig. Afterwards, the following steps are also included:
  • the information of the single virtual target vehicle is determined according to the risk weight coefficient and the second driving state information of each target vehicle;
  • the virtual driving scene includes a planned driving trajectory, and the planned driving trajectory is obtained according to the first driving state information of the own vehicle and the second driving state information of each target vehicle.
  • the information of the virtual target vehicle includes position information, speed information and acceleration information of the virtual target vehicle, and the position information, speed information and acceleration information of the virtual target vehicle satisfy the following formula:
  • i is the serial number of the target vehicle
  • ⁇ i is the risk weight coefficient of the target vehicle i
  • (x virtual , y virtual ) is the position information of the virtual target vehicle
  • a virtual driving scene can be displayed on a vehicle display screen, which can simulate actual road conditions, such as road condition information and information on various traffic signs or indicator lights. It also includes planning driving trajectories, and for multiple target vehicles within the preset range, it can be output as a single virtual target vehicle in the virtual driving scene. Realize more concise display effect and safer and more accurate path planning and driving control.
  • the prediction of the driving intention of the target vehicle and the adjustment of the risk weight coefficient in steps S403-step S408 of FIG. 4 or steps S503-step S508 of FIG. Handling efficiency and streamlining the flow of vehicle control.
  • the embodiment of this application does not make any limitation.
  • FIG. 6 is a schematic diagram of the comparison of the simulation results of the intelligent driving method and the single-target longitudinal control provided by the embodiment of the present application; where the abscissa is time, and the ordinate is acceleration.
  • the curve labeled 1 is a conventional single-target longitudinal control curve
  • the curve labeled 2 is the longitudinal control curve one generated by the method according to the embodiment of the application using the first set of calibration parameters (including various adjustment coefficients), and labeled 3
  • the curve of is the second longitudinal control curve generated by using the second set of calibration parameters (including various adjustment coefficients) according to the method of the embodiment of the present application.
  • Figure 7 is a schematic diagram of the composition of an intelligent driving device provided by the embodiment of the present application; it may include:
  • An acquisition unit 11 configured to acquire the first driving state information of the own vehicle and the second driving state information of a plurality of target vehicles, the target vehicles being located within a preset range;
  • the processing unit 12 is configured to determine a risk weight coefficient of each target vehicle according to the second driving state information of each target vehicle and the first driving state information, and the risk weight coefficient is used to adjust the driving state of the own vehicle.
  • processing unit 12 is specifically configured to:
  • the degree of influence satisfies a preset corresponding relationship with the second driving state information and the first driving state information of the target vehicle.
  • the first and/or second driving state information includes road condition information
  • the road condition information includes at least one of the following:
  • the road condition information is used to characterize the driving intention of each target vehicle.
  • the processing unit 12 when determining the driving intention of the first target vehicle among the various target vehicles according to the road condition information, is configured to:
  • the driving intention of the first target vehicle is determined according to the direction in which the degree of comprehensive influence decreases.
  • the driving intention of the first target vehicle affects the risk weight coefficient in the following manner:
  • the risk weight coefficient of the first target vehicle is increased
  • the risk weight coefficient of the first target vehicle is reduced.
  • processing unit 12 is further configured to:
  • the comprehensive acceleration of the own vehicle is calculated according to the expected acceleration of each target vehicle and the risk weight coefficient of each target vehicle.
  • processing unit 12 is further configured to:
  • the information of the single virtual target vehicle is determined according to the risk weight coefficient and the second driving state information of each target vehicle;
  • the virtual driving scene includes a planned driving trajectory, and the planned driving trajectory is obtained according to the first driving state information of the own vehicle and the second driving state information of each target vehicle.
  • the first driving state information includes the speed information of the own vehicle and acceleration information
  • the second driving state information includes the target vehicle's position information (xi , y i ), speed information and acceleration information
  • x′ i , y′ i are calculated according to the relative velocity and acceleration in the transverse direction and longitudinal direction as follows:
  • Parameters ⁇ x , ⁇ y , ⁇ x , ⁇ y can respectively adjust the degree of influence of lateral velocity, longitudinal velocity, lateral acceleration and longitudinal acceleration on normalization coefficient ⁇ x , ⁇ y , the larger ⁇ x , ⁇ y the smaller the influence range of potential energy .
  • the processing unit 12 calculates the degree of influence of the roadside line or lane line on the first target vehicle, satisfy the following formula:
  • the coefficient ⁇ lane is an adjustment coefficient for adjusting the degree of influence of the lane line/road; ⁇ lane is used for adjusting the size of the attenuation of the degree of influence; y lane is the lateral distance between the first target vehicle and the lane line/road;
  • the degree of influence of the passable obstacle on the first target vehicle satisfies the following formula:
  • the coefficient ⁇ obs is the adjustment coefficient for adjusting the influence degree of the passable obstacle
  • ⁇ x , ⁇ y are the adjustment coefficients respectively affecting the attenuation speed of the longitudinal and lateral influence degree
  • x obs , y obs are the passable obstacles object location.
  • the processing unit 12 estimates the degree of willingness of the first target vehicle to drive forward, which satisfies the following formula:
  • ⁇ attr is an adjustment coefficient for adjusting the degree of willingness.
  • the processing unit 12 calculates the first target vehicle according to the degree of influence of various information contained in the road condition information on the first target vehicle and the degree of willingness of the first target vehicle to drive forward.
  • the degree of comprehensive impact on the first target vehicle satisfies the following formula:
  • the processing unit 12 searches for a direction in which the degree of comprehensive influence is reduced by using a gradient descent method, which satisfies the following formula:
  • ⁇ is the gradient descent step size.
  • the processing unit 12 respectively calculates the expected acceleration of each target vehicle according to the car-following model, which satisfies the following formula:
  • v is the speed of the self-vehicle
  • ⁇ v is the speed difference between the self-vehicle and the vehicle in front of the current lane
  • v des is the expected speed
  • ⁇ is the calibration parameter
  • a max is the maximum acceleration
  • s * is the expected distance:
  • s 0 is the minimum distance of following and stopping; T is the following distance; b is the comfortable deceleration.
  • the processing unit 12 calculates the comprehensive acceleration of the own vehicle according to the expected acceleration of each target vehicle and the risk weight coefficient of each target vehicle, which satisfies the following formula:
  • a multi ⁇ i ⁇ a i ,
  • i is the serial number of the target vehicle
  • ⁇ i is the risk weight coefficient of the target vehicle i
  • a i is the expected acceleration of the target vehicle i.
  • the information of the virtual target vehicle includes position information, speed information and acceleration information of the virtual target vehicle, and the position information, speed information and acceleration information of the virtual target vehicle satisfy the following formula :
  • i is the serial number of the target vehicle
  • ⁇ i is the risk weight coefficient of the target vehicle i
  • (x virtual , y virtual ) is the position information of the virtual target vehicle
  • the calibration of various adjustment coefficients is determined according to external input parameters.
  • the calibration of various adjustment coefficients is determined according to the following objective function:
  • the objective function is used to determine the descending direction of the minimized influence on the actual driving direction of the ego vehicle at time t, wherein the direction vector of the ego vehicle at time t is Gradient descent direction of the degree of influence on the self-vehicle
  • Figure 8 is a schematic diagram of another intelligent driving device provided by the embodiment of the present application.
  • the intelligent driving device may include at least one processor 110 and a communication interface 120 .
  • the communication interface 120 is used to provide input or output of information, and the at least one processor 110 is used to execute programs or instructions, so that the processor 110 executes the method corresponding to the above Figure 2- Figure 5 where the intelligent driving device is located. The method performed by the device.
  • the communication interface 120 can acquire data collected by various sensors or cameras, such as the first driving state information of the own vehicle and the second state information of multiple target vehicles within a preset range, and can also output control data such as comprehensive acceleration or planned driving track to the controller.
  • the device may further include a memory, and the memory may be used to store instructions executed by the processor 110 , and further may be used to store data acquired by the communication interface 120 and data processed by the processor 110 .
  • the device may further include a bus, and various components in the device may be connected via the bus.
  • the device may be a chip including one or more processing circuits.
  • the one or more processing circuits are used to implement the method as described in any one of FIGS. 2-5 .
  • the communication interface 120 may be an input/output circuit or interface of a chip.
  • the device included in the device includes the above-mentioned chip, and the device may be a central control unit, etc., or a device independently configured on a vehicle for implementing the intelligent driving method in the embodiment of the present application, which is not limited in any way in the embodiment of the present application.
  • the function of the communication interface 120 may be considered to be implemented by a transceiver circuit or a dedicated chip for transceiver.
  • the processor 110 may be considered to be implemented by a dedicated processing chip, a processing circuit, a processor, or a general-purpose chip.
  • a general-purpose computer to implement the intelligent driving device provided in the embodiment of the present application.
  • the program codes for realizing the functions of the processor 110 and the communication interface 120 are stored in the memory, and the general-purpose processor implements the functions of the processor 110 and the communication interface 120 by executing the codes in the memory.
  • FIG. 8 only shows one processor. In an actual device, there may be multiple processors and memories.
  • a storage may also be called a storage medium or a storage device, etc., which is not limited in this embodiment of the present application.
  • the processor may be a central processing unit (Central Processing Unit, referred to as "CPU"), and the processor may also be other general processors, digital signal processors (DSP), application specific integrated circuits (ASIC ), off-the-shelf programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.
  • CPU Central Processing Unit
  • DSP digital signal processors
  • ASIC application specific integrated circuits
  • FPGA off-the-shelf programmable gate array
  • a general-purpose processor may be a microprocessor, or the processor may be any conventional processor, or the like.
  • the memory which can include read only memory and random access memory, provides instructions and data to the processor.
  • a portion of the memory may also include non-volatile random access memory.
  • the bus may also include a power bus, a control bus, and a status signal bus.
  • various buses are labeled as buses in the figures.
  • the embodiment of the present application also provides a vehicle control system.
  • vehicle control system For the relationship between various devices and the instruction flow, please refer to the description and description of the embodiment in Figure 1- Figure 5 description, which will not be repeated here.
  • the embodiment of the present application also provides a vehicle, and the relationship and instruction flow among the various devices included in the vehicle can be referred to the embodiments in Figure 1- Figure 5 The description and explanation of , will not be repeated here.
  • sequence numbers of the above-mentioned processes do not mean the order of execution, and the execution order of the processes should be determined by their functions and internal logic, and should not be used in the embodiments of the present application.
  • the implementation process constitutes any limitation.
  • the disclosed systems, devices and methods may be implemented in other ways.
  • the device embodiments described above are only illustrative.
  • the division of the units is only a logical function division. In actual implementation, there may be other division methods.
  • multiple units or components can be combined or May be integrated into another system, or some features may be ignored, or not implemented.
  • the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces, and the indirect coupling or communication connection of devices or units may be in electrical, mechanical or other forms.
  • all or part of them may be implemented by software, hardware, firmware or any combination thereof.
  • software When implemented using software, it may be implemented in whole or in part in the form of a computer program product.
  • the computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on the computer, the processes or functions according to the embodiments of the present application will be generated in whole or in part.
  • the computer can be a general purpose computer, a special purpose computer, a computer network, or other programmable devices.
  • the computer instructions may be stored in or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions may be transmitted from a website, computer, server or data center Transmission to another website site, computer, server, or data center by wired (eg, coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (eg, infrared, wireless, microwave, etc.).
  • the computer-readable storage medium may be any available medium that can be accessed by a computer, or a data storage device such as a server or a data center integrated with one or more available media.
  • the available medium may be a magnetic medium (for example, a floppy disk, a hard disk, or a magnetic tape), an optical medium (for example, DVD), or a semiconductor medium (for example, a Solid State Disk (SSD)).
  • SSD Solid State Disk

Abstract

一种智能驾驶方法、装置及车辆。方法包括：获取自车的第一行驶状态信息以及多个目标车辆的第二行驶状态信息，目标车辆位于预设范围内；根据各个目标车辆的第二行驶状态信息和第一行驶状态信息确定各个目标车辆的风险权重系数，风险权重系数用于调整自车的行驶状态。智能驾驶方法可以平滑的完成自车的纵向控制，提高驾驶的安全性和用户的舒适性。

Description

一种智能驾驶方法、装置及车辆 技术领域
本申请涉及智能驾驶技术领域，尤其涉及一种智能驾驶方法、装置及车辆。
背景技术
传统的自动驾驶纵向控制，只考虑单个目标车辆，这个目标车辆通常是自动驾驶的自车的当前车道最近车辆(Cloest In-Path Vehicle,CIPV)，这种控制方法可以称之为单目标纵向控制。CIPV信息作为自车控制模块的输入，需要上游在各种场景下实时、准确地选择CIPV，并且在必要时及时地进行CIPV的切换。
然而，CIPV的选择和切换面临许多困难，包括复杂场景下关键风险目标的选择和切换时机的把握。目前为止，即何时切换另外一个目标/释放当前目标，并无统一的标准和定义。由于控制模块只考虑上游输出的单个目标车辆，在上述CIPV的切换场景中，很有可能导致自车先加速、再减速的不舒适和不安全过程。CIPV切换过早容易造成自车误判减速，切换太晚又容易导致自车急刹。而且在城市拥堵路况等复杂场景下，要从前方许多有风险的目标中选出最具代表性一个车辆来实现平滑、安全的控制任务是很困难的。因此，无论CIPV选择算法的性能有多好，在复杂场景，仅考虑单个目标的纵向控制远远不够。
发明内容
本申请实施例所要解决的技术问题在于，提供一种智能驾驶方法、装置及车辆，以平滑的完成自车的纵向控制，提高驾驶的安全性和用户的舒适性。
第一方面，本申请的实施例提供了一种智能驾驶方法，可包括：
获取自车的第一行驶状态信息以及多个目标车辆的第二行驶状态信息，所述目标车辆位于预设范围内；
根据各个目标车辆的第二行驶状态信息和所述第一行驶状态信息确定各个目标车辆的风险权重系数，所述风险权重系数用于调整所述自车的行驶状态。
通过获取自车的第一行驶状态信息和预设范围内的多个目标车辆的第二行驶状态信息，并对预设范围内的多个目标车辆进行评估，得到多个目标车辆对自车的风险权重系数，从而可以根据多个目标车辆的风险权重系数来控制自车的行驶状态，由于综合考虑的预设范围内的多个目标车辆的信息，因此了提升了驾驶的安全性，且避免了单个目标车辆在切换和释放的过程中带来的急加速和急减速等情况，因此也提升了用户的舒适性。
在一种可能的实现方式中，所述根据各个目标车辆的第二行驶状态信息和所述第一行驶状态信息确定各个目标车辆的风险权重系数，包括：
根据所述各个目标车辆对所述自车的影响程度确定各个目标车辆的风险权重系数；
其中，所述影响程度与所述目标车辆的第二行驶状态信息和所述第一行驶状态信息之间满足预先设置的对应关系。
通过各个目标车辆对自车的影响程度来评估各个目标车辆的风险权重系数，可以提升综合控制的全面性和准确性。
在一种可能的实现方式中，所述第一和/或第二行驶状态信息包含道路状况信息，所述道 路状况信息包括以下至少一种：
所述预设范围内的其他车辆信息、路沿信息、车道线信息或可通行障碍物信息。
可以通过道路状况信息中的路沿信息和车道线信息等确定目标车辆的位置信息，进而更加准确完成风险权重系数的确定。且综合考虑路况信息可以提升自车控制的安全性。
在一种可能的实现方式中，所述道路状况信息用于表征所述各个目标车辆的行驶意图。
在一种可能的实现方式中，所述方法还包括：
分别计算所述道路状况信息中包含的各项信息对第一目标车辆的影响程度；
估计所述第一目标车辆向前行驶的意愿程度；
根据所述道路状况信息中包含的各项信息对第一目标车辆的影响程度以及所述第一目标车辆向前行驶的意愿程度，计算所述第一目标车辆受到的综合影响程度；
通过梯度下降方法搜索所述综合影响程度降低的方向；
根据所述综合影响程度降低的方向确定所述第一目标车辆的行驶意图。
通过提前估计目标车辆的行驶意图，并根据目标车辆的行驶意图进一步的调整风险权重系数。可以更进一步的提升风险权重系数的准确性，进而提升驾驶的安全性和控制的前瞻性。
在一种可能的实现方式中，所述第一目标车辆的行驶意图通过以下方式影响所述风险权重系数：
若所述第一目标车辆的行驶意图为驶入所述自车当前所处车道，则所述第一目标车辆的风险权重系数提高；或者
若所述第一目标车辆的形式意图为驶出所述自车当前所处车道，则所述第一目标车辆的风险权重系数降低。
通过这样的调整规则可以提升风险权重系数的准确性。
在一种可能的实现方式中，所述方法还包括：
根据跟驰模型分别计算各个目标车辆的期望加速度；
根据各个目标车辆的期望加速度以及各个目标车辆的风险权重系数计算所述自车的综合加速度。
通过输出综合加速度，控制器便可以根据综合加速度来控制自车的行驶状态。由于该综合加速度中考虑了多个目标车辆的期望加速度和风险权重系数，因此可以实现对自车更加准确且安全的控制。
在一种可能的实现方式中，所述方法还包括：
输出规划的虚拟行驶场景以及投影到所述虚拟行驶场景的单个虚拟目标车辆的信息，所述单个虚拟目标车辆的信息根据所述各个目标车辆的风险权重系数和第二行驶状态信息确定；
所述虚拟行驶场景包含规划行驶轨迹，所述规划行驶轨迹是根据所述自车的第一行驶状态信息以及各个目标车辆的第二行驶状态信息规划得到的。
通过输出一个虚拟目标场景和规划形式轨迹，可以实现更加简洁的显示效果以及更加安全且准确的路径规划和行驶控制。
在一种可能的实现方式中，所述第一行驶状态信息包括所述自车的速度信息
Figure PCTCN2022074581-appb-000001
和加速度信息
Figure PCTCN2022074581-appb-000002
所述第二行驶状态信息包括目标车辆的位置信息(x i,y i)、速度信息
Figure PCTCN2022074581-appb-000003
和加速度信息
Figure PCTCN2022074581-appb-000004
所述各个目标车辆对所述自车的影响程度，满足以下公式：
Figure PCTCN2022074581-appb-000005
其中，K为欧氏距离且K＝‖x′ i,y′ i‖ 2，x′ i,y′ i为规范化后的位置信息，分别用于表征目标车辆i与所述自车的横向距离和纵向距离，i为目标车辆的序号，系数ω car为用于调节目标车辆的对所述自车的影响程度的调节系数；系数α用于调节目标车辆影响范围大小；
其中，x′ i,y′ i根据横向和纵向的相对速度和加速度计算如下：
Figure PCTCN2022074581-appb-000006
Figure PCTCN2022074581-appb-000007
参数β x,β y,γ x,γ y可分别调节横向速度、纵向速度、横向加速度和纵向加速度对规范化系数∈ x,∈ y的影响程度，∈ x,∈ y越大势能影响范围越小。
在一种可能的实现方式中，所述分别计算所述道路状况信息中包含的各项信息对第一目标车辆的影响程度，包括：
当所述道路状况信息中包含路沿信息或车道线信息时，所述路沿线或车道线对所述第一目标车辆的影响程度，满足以下公式：
Figure PCTCN2022074581-appb-000008
其中，系数ω lane为用于调整车道线/路沿线的影响程度的调节系数；δ lane用于调整影响程度衰减的大小；y lane为第一目标车辆与车道线/路沿线的横向距离；
当所述道路状况信息中包含可通行障碍物信息时，所述可通行障碍物对第一目标车辆的影响程度，满足以下公式：
Figure PCTCN2022074581-appb-000009
其中，系数ω obs为调整所述可通行障碍物的影响程度的调节系数；δ x,δ y为分别影响纵向和横向影响程度衰减速度的调节系数；x obs,y obs为所述可通行障碍的物位置。
在一种可能的实现方式中，所述估计所述第一目标车辆向前行驶的意愿程度，满足以下公式：
U attr(x)＝-ω attr·x
其中，ω attr为用于调整意愿程度的大小的调节系数。
在一种可能的实现方式中，所述根据所述道路状况信息中包含的各项信息对第一目标车辆的影响程度以及所述第一目标车辆向前行驶的意愿程度，计算所述第一目标车辆受到的综合影响程度，满足以下公式：
Figure PCTCN2022074581-appb-000010
其中，
Figure PCTCN2022074581-appb-000011
表示所述其他车辆对第一目标车辆i的影响程度。
在一种可能的实现方式中，所述通过梯度下降方法搜索所述综合影响程度降低的方向，满足以下公式：
Figure PCTCN2022074581-appb-000012
其中，η为梯度下降步长。
在一种可能的实现方式中，所述根据跟驰模型分别计算各个目标车辆的期望加速度，满足以下公式：
Figure PCTCN2022074581-appb-000013
其中，v为所述自车的速度；Δv为所述自车和当前车道的前车的速度差；v des为期望速度；σ为标定参数；a max为最大加速度；s为当前时刻所述自车和所述前车的距离；s *为期望距离：
Figure PCTCN2022074581-appb-000014
其中，s 0为跟停最小距离；T为跟车时距；b为舒适减速度。
在一种可能的实现方式中，所述根据各个目标车辆的期望加速度以及各个目标车辆的风
险权重系数计算所述自车的综合加速度，满足以下公式：
a multi＝∑ω i·a i，
其中，i为目标车辆的序号，ω i为目标车辆i的风险权重系数，a i为目标车辆i的期望加速度。
在一种可能的实现方式中，所述虚拟目标车辆的信息包括所述虚拟目标车辆的位置信息、速度信息和加速度信息，所述虚拟目标车辆的位置信息、速度信息和加速度信息，满足以下公式：
(x virtual,y virtual)＝(∑ω i·x i,∑ω i·y i)，
Figure PCTCN2022074581-appb-000015
Figure PCTCN2022074581-appb-000016
其中，i为目标车辆的序号，ω i为目标车辆i的风险权重系数，(x virtual,y virtual)为所述虚拟目标车辆的位置信息，
Figure PCTCN2022074581-appb-000017
为所述虚拟目标车辆的速度信息，
Figure PCTCN2022074581-appb-000018
为所述虚拟目标车辆的加速度信息。
在一种可能的实现方式中，各种调节系数的标定根据来自外部的输入参数确定。
在一种可能的实现方式中，各种调节系数的标定根据如下目标函数确定：
Figure PCTCN2022074581-appb-000019
所述目标函数用于确定t时刻对于自车实际驾驶方向的最小化影响程度下降方向，其中，在t时刻所述自车方向向量为
Figure PCTCN2022074581-appb-000020
对所述自车影响程度梯度下降方向
Figure PCTCN2022074581-appb-000021
第二方面，本申请的实施例提供了一种智能驾驶装置，可包括：
获取单元，用于获取自车的第一行驶状态信息以及多个目标车辆的第二行驶状态信息，所述目标车辆位于预设范围内；
处理单元，用于根据各个目标车辆的第二行驶状态信息和所述第一行驶状态信息确定各个目标车辆的风险权重系数，所述风险权重系数用于调整所述自车的行驶状态。
在一种可能的实现方式中，所述处理单元具体用于：
根据所述各个目标车辆对所述自车的影响程度确定各个目标车辆的风险权重系数；
其中，所述影响程度与所述目标车辆的第二行驶状态信息和所述第一行驶状态信息之间满足预先设置的对应关系。
在一种可能的实现方式中，所述第一和/或第二行驶状态信息包含道路状况信息，所述道路状况信息包括以下至少一种：
所述预设范围内的其他车辆信息、路沿信息、车道线信息或可通行障碍物信息。
在一种可能的实现方式中，所述道路状况信息用于表征所述各个目标车辆的行驶意图。
在一种可能的实现方式中，当根据所述道路状况信息确定所述各个目标车辆中的第一目标车辆的行驶意图时，所述处理单元用于：
分别计算所述道路状况信息中包含的各项信息对所述第一目标车辆的影响程度；
估计所述第一目标车辆向前行驶的意愿程度；
根据所述道路状况信息中包含的各项信息对第一目标车辆的影响程度以及所述第一目标车辆向前行驶的意愿程度，计算所述第一目标车辆受到的综合影响程度；
通过梯度下降方法搜索所述综合影响程度降低的方向；
根据所述综合影响程度降低的方向确定所述第一目标车辆的行驶意图。
在一种可能的实现方式中，所述第一目标车辆的行驶意图通过以下方式影响所述风险权重系数：
若所述第一目标车辆的行驶意图为驶入所述自车当前所处车道，则所述第一目标车辆的风险权重系数提高；或者
若所述第一目标车辆的形式意图为驶出所述自车当前所处车道，则所述第一目标车辆的风险权重系数降低。
在一种可能的实现方式中，所述处理单元还用于：
根据跟驰模型分别计算各个目标车辆的期望加速度；
根据各个目标车辆的期望加速度以及各个目标车辆的风险权重系数计算所述自车的综合加速度。
在一种可能的实现方式中，所述处理单元还用于：
输出规划的虚拟行驶场景以及投影到所述虚拟行驶场景的单个虚拟目标车辆的信息，所述单个虚拟目标车辆的信息根据所述各个目标车辆的风险权重系数和第二行驶状态信息确定；
所述虚拟行驶场景包含规划行驶轨迹，所述规划行驶轨迹是根据所述自车的第一行驶状态信息以及各个目标车辆的第二行驶状态信息规划得到的。
在一种可能的实现方式中，所述第一行驶状态信息包括所述自车的速度信息
Figure PCTCN2022074581-appb-000022
和加速度信息
Figure PCTCN2022074581-appb-000023
所述第二行驶状态信息包括目标车辆的位置信息(x i,y i)、速度信息
Figure PCTCN2022074581-appb-000024
和加速度信息
Figure PCTCN2022074581-appb-000025
所述各个目标车辆对所述自车的影响程度，满足以下公式：
Figure PCTCN2022074581-appb-000026
其中，K为欧氏距离且K＝‖x′ i,y′ i‖ 2，x′ i,y′ i为规范化后的位置信息，分别用于表征目标车辆i与所述自车的横向距离和纵向距离，i为目标车辆的序号，系数ω car为用于调节目标车辆的对所述自车的影响程度的调节系数；系数α用于调节目标车辆影响范围大小；
其中，x′ i,y′ i根据横向和纵向的相对速度和加速度计算如下：
Figure PCTCN2022074581-appb-000027
Figure PCTCN2022074581-appb-000028
参数β x,β y,γ x,γ y可分别调节横向速度、纵向速度、横向加速度和纵向加速度对规范化系数∈ x,∈ y的影响程度，∈ x,∈ y越大势能影响范围越小。
在一种可能的实现方式中，当所述道路状况信息中包含路沿信息或车道线信息时，所述处理单元计算所述路沿线或车道线对所述第一目标车辆的影响程度，满足以下公式：
Figure PCTCN2022074581-appb-000029
其中，系数ω lane为用于调整车道线/路沿线的影响程度的调节系数；δ lane用于调整影响程度衰减的大小；y lane为第一目标车辆与车道线/路沿线的横向距离；
当所述道路状况信息中包含可通行障碍物信息时，所述可通行障碍物对第一目标车辆的影响程度，满足以下公式：
Figure PCTCN2022074581-appb-000030
其中，系数ω obs为调整所述可通行障碍物的影响程度的调节系数；δ x,δ y为分别影响纵向和横向影响程度衰减速度的调节系数；x obs,y obs为所述可通行障碍的物位置。
在一种可能的实现方式中，所述处理单元估计所述第一目标车辆向前行驶的意愿程度，满足以下公式：
U attr(x)＝-ω attr·x
其中，ω attr为用于调整意愿程度的大小的调节系数。
在一种可能的实现方式中，所述处理单元根据所述道路状况信息中包含的各项信息对第一目标车辆的影响程度以及所述第一目标车辆向前行驶的意愿程度，计算所述第一目标车辆受到的综合影响程度，满足以下公式：
Figure PCTCN2022074581-appb-000031
其中，
Figure PCTCN2022074581-appb-000032
表示所述其他车辆对第一目标车辆i的影响程度。
在一种可能的实现方式中，所述处理单元通过梯度下降方法搜索所述综合影响程度降低的方向，满足以下公式：
Figure PCTCN2022074581-appb-000033
其中，η为梯度下降步长。
在一种可能的实现方式中，所述处理单元根据跟驰模型分别计算各个目标车辆的期望加速度，满足以下公式：
Figure PCTCN2022074581-appb-000034
其中，v为所述自车的速度；Δv为所述自车和当前车道的前车的速度差；v des为期望速度；σ为标定参数；a max为最大加速度；s为当前时刻所述自车和所述前车的距离；s *为期望距离：
Figure PCTCN2022074581-appb-000035
其中，s 0为跟停最小距离；T为跟车时距；b为舒适减速度。
在一种可能的实现方式中，所述处理单元根据各个目标车辆的期望加速度以及各个目标车辆的风险权重系数计算所述自车的综合加速度，满足以下公式：
a multi＝∑ω i·a i，
其中，i为目标车辆的序号，ω i为目标车辆i的风险权重系数，a i为目标车辆i的期望加速度。
在一种可能的实现方式中，所述虚拟目标车辆的信息包括所述虚拟目标车辆的位置信息、速度信息和加速度信息，所述虚拟目标车辆的位置信息、速度信息和加速度信息，满足以下公式：
(x virtual,y virtual)＝∑ω i·x i,∑ω i·y i)，
Figure PCTCN2022074581-appb-000036
Figure PCTCN2022074581-appb-000037
其中，i为目标车辆的序号，ω i为目标车辆i的风险权重系数，(x virtual,y virtual)为所述虚拟目标车辆的位置信息，
Figure PCTCN2022074581-appb-000038
为所述虚拟目标车辆的速度信息，
Figure PCTCN2022074581-appb-000039
为所述虚拟目标车辆的加速度信息。
在一种可能的实现方式中，各种调节系数的标定根据来自外部的输入参数确定。
在一种可能的实现方式中，各种调节系数的标定根据如下目标函数确定：
Figure PCTCN2022074581-appb-000040
所述目标函数用于确定t时刻对于自车实际驾驶方向的最小化影响程度下降方向，其中，在t时刻所述自车方向向量为
Figure PCTCN2022074581-appb-000041
对所述自车影响程度梯度下降方向
Figure PCTCN2022074581-appb-000042
第三方面，本申请的实施例提供了一种智能驾驶装置，可包括：
处理器、存储器和总线，所述处理器和存储器通过总线连接，其中，所述存储器用于存储一组程序代码，所述处理器用于调用所述存储器中存储的程序代码，执行本申请实施例第一方面或第一方面任一实现方式中的方法。
第四方面，本申请的实施例提供了一种计算机可读存储介质，所述计算机可读存储介质中存储有指令，当其在计算机上运行时，实现上述第一方面或第一方面任一实现方式所述的方法。
第五方面，提供了一种计算机程序产品，所述计算机程序产品包括：计算机程序代码，当所述计算机程序代码在计算机上运行时，使得计算机执行上述第一方面或第一方面任一实现方式所述的方法。
第六方面，本申请实施例提供了一种智能驾驶装置，该装置可以是车辆中的芯片或片上系统，该装置包括处理器，所述处理器与存储器耦合，该存储器用于存储计算机程序或指令，该处理器用于执行该存储器中的计算机程序或指令，使得该装置执行如第一方面或第一方面任一实现方式所述的方法，可选的，该装置还包括该存储器。
第七方面，本申请的实施例提供了一种车辆，可包括：
如本申请实施例第二方面或第二方面任一实现方式中的装置。
附图说明
为了更清楚地说明本申请实施例或背景技术中的技术方案，下面将对本申请实施例或背景技术中所需要使用的附图进行说明。
图1为本申请实施例提供的一种车辆的控制系统的架构示意图；
图2为本申请实施例提供的一种智能驾驶方法的流程示意图；
图3为本申请实施例提供的另一种智能驾驶方法的流程示意图；
图4为本申请实施例提供的又一种智能驾驶方法的流程示意图；
图5为本申请实施例提供的又一种智能驾驶方法的流程示意图；
图6为本申请实施例提供的智能驾驶方法与单目标纵向控制的仿真结果对比示意图；
图7为本申请实施例提供的一种智能驾驶装置的组成示意图；
图8为本申请实施例提供的另一种智能驾驶装置的组成示意图。
具体实施方式
下面结合本申请实施例中的附图对本申请的实施例进行描述。
本申请中提及的“包括”和“具有”以及它们任何变形，意图在于覆盖不排他的包含。例如，包含了一系列步骤或单元的过程、方法、系统、产品或设备并不限定于已列出的步骤或单元，而是可选地还包括没有列出的步骤或单元，或可选地还包括对于这些过程、方法、产品或设备固有的其它步骤或单元。
根据工信部公示的《汽车驾驶自动化分级》，国内的自动驾驶可划分为以下六个等级：
L0级驾驶自动化(应急辅助):驾驶自动化系统不能持续执行动态驾驶任务中的车辆横向或纵向运动控制，但具备持续执行动态驾驶任务中的部分目标和事件探测与响应的能力。
L1级驾驶自动化(部分驾驶辅助):驾驶自动化系统在其设计运行条件内持续地执行动态驾驶任务中的车辆横向或纵向运动控制，且具备与所执行的车辆横向或纵向运动控制相适应的部分目标和事件探测与响应的能力。
L2级驾驶自动化(组合驾驶辅助):驾驶自动化系统在其设计运行条件内持续地执行动态驾驶任务中的车辆横向和纵向运动控制，且具备与所执行的车辆横向和纵向运动控制相适应的部分目标和事件探测与响应的能力。
L3级驾驶自动化(有条件自动驾驶):驾驶自动化系统在其设计运行条件内持续地执行全部动态驾驶任务。
L4级驾驶自动化(高度自动驾驶):驾驶自动化系统在其设计运行条件内持续地执行全部动态驾驶任务和执行动态驾驶任务接管。
L5级驾驶自动化(完全自动驾驶):驾驶自动化系统在任何可行驶条件下持续地执行全部动态驾驶任务和执行动态驾驶任务接管。
本申请实施例中智能驾驶方法可以适用于国内L1/L2级别的自动驾驶，还可以适用于国内L2级别以上的自动驾驶场景。且对于美国机动车工程师学会(Society of Automotive Engineers，SAE)制定的L1-L5级别的自动驾驶同样适用。
由于在实际的驾驶场景中，对于自车而言，道路上通常会存在多台其他车辆，以及可能存在的一些道路状况信息如路沿、车道线、障碍物等。这些物体都会对自车产生一定的影响，本申请中描述的影响程度可通过某物体对自车的影响进行量化得到的数据进行表征。通常情况下，对自车的影响越大，对应数据的数值越大，对自车的影响越小，对应数据的数值越小。
图1为本申请实施例提供的一种车辆的控制系统的架构示意图。可包括：
智能驾驶装置10、雷达20、摄像头30、惯性测量单元(Inertial Measurement Unit，IMU)40和控制器50。
其中，智能驾驶装置10用于通过自车搭载的各种传感器以感知自车的情况以及周边环境的情况，例如，可以通过IMU获取自车的行驶状态信息如速度信息和加速度信息等；可以通过雷达20感知周边环境的车辆、障碍物等；也可以通过摄像头30拍摄道路状况信息如周边车辆信息、车道线信息、路沿信息等。在本申请实施例中，智能驾驶装置10可以根据获取到的预设范围内多台目标车辆的第二行驶状态信息(包括位置、速度和加速度等信息)来最终确定自车的行驶状态。
雷达20的类型可以是超声波雷达、微波雷达或激光雷达等。自车上搭载的雷达数目可以是一个或多个。可用于发现障碍物、预测碰撞、自适应巡航控制等。雷达20可以被安装在自车的前表面的中心部分及角部、后表面的中心部分以及角部等，可以在雷达20的前方区域的预定角度范围内发射电磁波或激光，可以接收从位于自车附近的周边对象反射回来的回波，从而可以检测自车和各个周边对象之间的角度、距离、速度、加速度等，并将这些信息发送 至智能驾驶装置10。
摄像头30的类型可以是红外线摄像头、互补金属氧化物半导体(Complementary Metal Oxide Semiconductor，CMOS)摄像头、电荷耦合器件(charge coupled device，CCD)摄像头或激光摄像头等。自车上搭载的摄像头数目可以是一个或多个。可用于拍摄自车周边的环境和车辆等。摄像头30同样可以被安装在自车的前表面的中心部分及角部、后表面的中心部分以及角部等，也可以被安装在主车辆的挡风玻璃的上端，可以在相对于摄像头30的前方区域的预定角度和预定距离范围内感测和投射各种类型的光(例如红外光、可见光等)，可以获取自车周围的外部对象的图像，并且可以将所获取的图像发送至智能驾驶装置10。
智能驾驶装置10可以包括获取单元11和处理单元12，获取单元11用于从各种传感器获取数据，处理器用于处理获取到的数据。此外，智能驾驶装置10还可以包括诸如只读存储器(ROM)或随机存取存储器(RAM)的存储器(例如，DB)，可以存储获取到的数据，以及处理单元12需要执行的程序代码，还可以存储处理单元12处理获取到的数据之后生成的新数据。例如，在本申请实施例中，获取单元11可以通过IMU40获取自车的第一行驶状态信息，还可以通过雷达20和摄像头30等传感器获取预设范围内的目标车辆的第二行驶状态信息；处理单元12可以根据获取单元11获取的信息将确定各个目标车辆对自车的影响程度，并分别确定各个目标车辆相对于自车的风险权重系数，从而输出更加平滑准确的控制信息如综合加速度或行驶轨迹等给控制器50。
控制器50可以是比例-积分-微分(Proportion Integration Differentiation,PID)控制器或其他类型的控制器。控制器50可以包括诸如只读存储器(ROM)或随机存取存储器(RAM)的存储器(例如，DB)，可以存储各种控制数据和控制程序代码，并且可以进一步包括诸如CPU之类的处理器，使得控制器50可以执行各种控制程序。接收智能驾驶装置10输出的控制信息，根据控制信息实现对自车行驶状态的调整。如根据智能驾驶装置输出的综合加速度或行驶轨迹，控制自车的速度和/或加速度。
需要说明的是，除了上述的雷达20、摄像头30和IMU40之外，自车上还可以搭载诸如光探测和测距传感器等任何可用于提升自车感应能力的传感器。本申请实施例不作任何限定。
下面结合图2-图4对本申请智能驾驶方法进行详细描述。
请参见图2，图2为本申请实施例提供的一种智能驾驶方法的流程示意图；可包括如下步骤：
S201.获取自车的第一行驶状态信息以及多个目标车辆的第二行驶状态信息。
参照图1所示，自车可以通过IMU获取自车的第一行驶状态信息如速度信息和加速度信息等；还可以通过雷达20和摄像头30等传感器获取预设范围内的多个目标车辆的第二行驶状态信息如位置信息、速度信息、加速度信息和道路状况信息等。
其中，所述目标车辆位于预设范围内且数量大于1个。该预设范围可以由自车的控制器根据外部输入的数据设定，也可以由远端的服务器进行设定，还可以由用户根据自身的需求进行设定或调节。例如，在视线受阻的雨雾天气或者当前用户非常注重安全的情况下，可以将预设范围设置更大一些，而在视线正常的天气或者当前用户希望提升自车处理数据效率的情况下，可以将预设范围设置更小一些。
自车的第一行驶状态信息可以包括自车的速度信息和加速度信息，或者还可以包括道路状况信息；第二行驶状态信息根据目标车辆的数目对应地存在多个。每一个目标车辆的第二行驶状态信息可以包括该目标车辆的位置信息，速度信息和加速度信息，或者还可以包括道路状况信息。所述道路状况信息包括以下至少一种：
所述预设范围内的其他车辆信息、路沿信息、车道线信息或可通行障碍物信息。
其中，位置信息可以根据自车获取的路沿信息和/或车道线信息等来确定。
S202.根据各个目标车辆的第二行驶状态信息和所述第一行驶状态信息确定各个目标车辆的风险权重系数。
其中，所述风险权重系数用于调整所述自车的行驶状态。
由于预设范围内的各个目标车辆均会对自车形成不同程度的影响，例如，自车当前所处车道前方的目标车辆是否加减速，是否变道等会对自车造成影响；自车左侧或右侧的车道是否变道会对自车造成影响等。
因此，可以根据所述各个目标车辆对所述自车的影响程度确定各个目标车辆的风险权重系数；
其中，所述影响程度与所述目标车辆的第二行驶状态信息和所述第一行驶状态信息之间满足预先设置的对应关系。
一种设计中，所述各个目标车辆对所述自车的影响程度，满足以下公式：
Figure PCTCN2022074581-appb-000043
其中，K为欧氏距离且K＝‖x′ i,y′ i‖ 2，x′ i,y′ i为规范化后的位置信息，分别用于表征目标车辆i与所述自车的横向距离和纵向距离，i为目标车辆的序号，系数ω car为用于调节目标车辆的对所述自车的影响程度的调节系数；系数α用于调节目标车辆影响范围大小；
其中，x′ i,y′ i根据横向和纵向的相对速度和加速度计算如下：
Figure PCTCN2022074581-appb-000044
Figure PCTCN2022074581-appb-000045
参数β x,β y,γ x,γ y可分别调节横向速度、纵向速度、横向加速度和纵向加速度对规范化系数∈ x,∈ y的影响程度，∈ x,∈ y越大势能影响范围越小。
当计算得到各个目标车辆对自车的影响程度的数据之后，可以对所有数据进行归一化处理，从而得到各个目标车辆的风险权重系数。
例如，预设范围内存在4台目标车辆，对自车的影响程度计算得到的数据分别为80,60，40,20。则进行归一化处理后得到的风险权重系数分别为40％,30％,20％,10％。
在本申请实施例中，通过获取自车的第一行驶状态信息和预设范围内的多个目标车辆的第二行驶状态信息，并对预设范围内的多个目标车辆进行评估，得到多个目标车辆对自车的风险权重系数，从而可以根据多个目标车辆的风险权重系数来控制自车的行驶状态，由于综合考虑的预设范围内的多个目标车辆的信息，因此了提升了驾驶的安全性，且避免了单个目标车辆在切换和释放的过程中带来的急加速和急减速等情况，因此也提升了用户的舒适性。
由于目标车辆的行驶状态可能会发生变化，例如加减速或变道等，道路状况也可能发生变化，例如前方出现不可变道的实线车道线，或路沿变窄，或出现可通行障碍物等。因此，可以道路状况信息可用于表征所述各个目标车辆的行驶意图。通过提前估计目标车辆的行驶意图，可以根据目标车辆的行驶意图进一步的调整风险权重系数。
请参见图3，图3为本申请实施例提供的另一种智能驾驶方法的流程示意图；其中，步骤S301-步骤S302和图2中的步骤S201-步骤S202相同，此处不再赘述，在步骤S302之后，还包括如下步骤：
S303.分别计算所述道路状况信息中包含的各项信息对第一目标车辆的影响程度。
可选地，当所述道路状况信息中包含预设范围内的其他车辆时，可以参照步骤S202中计算目标车辆对自车影响程度的方式来计算其他车辆对第一目标车辆的影响程度。
需要说明的是，在步骤S202中计算目标车辆对自车的影响程度时，可以考虑自车前方的目标车辆来进行计算。在步骤S303中计算其他车辆对第一目标车辆的影响程度时，可以同时考虑第一目标车辆前方的其他车辆以及后方的其他车辆来进行计算。
一种设计中，当所述道路状况信息中包含路沿信息或车道线信息时，所述路沿线或车道线对所述第一目标车辆的影响程度，满足以下公式：
Figure PCTCN2022074581-appb-000046
其中，系数ω lane为用于调整车道线/路沿线的影响程度的调节系数；δ lane用于调整影响程度衰减的大小；y lane为第一目标车辆与车道线/路沿线的横向距离；
当所述道路状况信息中包含可通行障碍物信息时，所述可通行障碍物对第一目标车辆的影响程度，满足以下公式：
Figure PCTCN2022074581-appb-000047
其中，系数ω obs为调整所述可通行障碍物的影响程度的调节系数；δ x,δ y为分别影响纵向和横向影响程度衰减速度的调节系数；x obs,y obs为所述可通行障碍的物位置。
S304.估计所述第一目标车辆向前行驶的意愿程度。
所述意愿程度用于表征第一目标车辆向前行驶的惯性。例如，第一目标车辆持续一段时间内的速度和加速度均未发生变化或未发生较大变化，则可确定第一目标车辆向前行驶的意愿强度较强；若第一目标车辆的横向加速度发生变化，则可能存在变道或拐弯或掉头的可能性，向前行驶的意愿程度较低。
一种设计中，估计所述第一目标车辆向前行驶的意愿程度，满足以下公式：
U attr(x)＝-ω attr·x
其中，ω attr为用于调整意愿程度的大小的调节系数。
S305.根据所述道路状况信息中包含的各项信息对第一目标车辆的影响程度以及所述第一目标车辆向前行驶的意愿程度，计算所述第一目标车辆受到的综合影响程度。
一种设计中，计算所述第一目标车辆受到的综合影响程度，满足以下公式：
Figure PCTCN2022074581-appb-000048
其中，
Figure PCTCN2022074581-appb-000049
表示所述其他车辆对第一目标车辆i的影响程度。
S306.通过梯度下降方法搜索所述综合影响程度降低的方向。
梯度下降方法(gradient descent)常用于机器学习和人工智能当中用来递归性地逼近最小偏差模型。其计算过程包含沿梯度下降的方向通过迭代找到目标函数的最小值或者收敛到最小值。
一种设计中，所述通过梯度下降方法搜索所述综合影响程度降低的方向，满足以下公式：
Figure PCTCN2022074581-appb-000050
其中，η为梯度下降步长。
S307.根据所述综合影响程度降低的方向确定所述第一目标车辆的行驶意图。
通过梯度下降方法得到综合程度降低的方向就可以预估第一目标车辆未来可能的行驶方向。进而确定第一目标车辆的行驶意图。例如，综合影响程度降低的方向对应第一目标车辆可能向右行驶，第一目标车辆的行驶意图为变道并进入自车所在车道，则此时第一目标车辆对自车的影响将会变大。若综合影响程度降低的方向对应第一目标车辆可能向左行驶，第一目标车辆的行驶意图为变道并离开自车所在车道，则此时第一目标车辆对自车的影响将会变小。若综合影响程度降低的方向对应第一目标车辆可能保持前行，第一目标车辆的行驶意图为保持当前行驶状态在原所在车道行驶，则此时第一目标车辆对自车的影响将不变。
S308.根据所述第一目标车辆的行驶意图调整所述第一目标车辆的风险权重系数。
可选地，若所述第一目标车辆的行驶意图为驶入所述自车当前所处车道，则所述第一目标车辆的风险权重系数提高；或者
若所述第一目标车辆的形式意图为驶出所述自车当前所处车道，则所述第一目标车辆的风险权重系数降低。
当调整风险权重系数后，再一次对预设范围内的目标车辆的风险权重系数进行归一化处理。
可选地，本申请实施例中的各种调节系数如ω lane、ω car、ω obs、ω attr、δ lane、δ x、δ y等的标定根据来自外部的输入参数确定。
或者，本申请实施例中的各种调节系数的标定根据如下目标函数确定：
Figure PCTCN2022074581-appb-000051
所述目标函数用于确定t时刻对于自车实际驾驶方向的最小化影响程度下降方向，其中，在t时刻所述自车方向向量为
Figure PCTCN2022074581-appb-000052
对所述自车影响程度梯度下降方向
Figure PCTCN2022074581-appb-000053
若车辆方向与梯度下降方向一致，则
Figure PCTCN2022074581-appb-000054
若方向相反则cosθ＝-1。通过求解上述最优化问题，可得到参数ω car,ω lane,ω obs,ω attr,δ lane,δ x,δ y。
在本申请实施例中，通过对目标车辆的行驶意图进行预测，从而可以更进一步的对目标车辆的风险权重系数进行精细化的调整，得到更符合未来综合路况的信息，进一步提升了驾驶的安全性和控制的前瞻性。
针对L1/L2级别的自动驾驶，本申请实施例可输出综合加速度至控制器以便控制器控制自车的行驶状态。请参见图4，图4为本申请实施例提供的又一种智能驾驶方法的流程示意图；步骤S401-步骤S408与图3中的步骤S301-步骤S308相同，此处不再赘述，在步骤S308之后，还包括如下步骤：
S409.根据跟驰模型分别计算各个目标车辆的期望加速度。
车辆跟驰(Car Following,CF)行为是最基本的微观驾驶行为，描述了在限制超车的单行道上行驶车队中相邻两车之间的相互作用。跟驰模型是运用动力学的方法来研究前导车(Leading Vehicle，LV)运动状态变化所引起跟驰车(Following Vehicle，FV)的相应行为，通过分析各车辆逐一跟驰的方式来理解单车道交通流特性，从而在驾驶人微观行为与交通宏观现象之间架起一座桥梁。本申请实施例的方法不限定跟驰模型的具体形式或版本，可适用于当前已存在或未来更先进的各种版本的跟驰模型。
一种设计中，根据跟驰模型分别计算各个目标车辆的期望加速度，满足以下公式：
Figure PCTCN2022074581-appb-000055
其中，v为所述自车的速度；Δv为所述自车和当前车道的前车的速度差；v des为期望速度；σ为标定参数；a max为最大加速度；s为当前时刻所述自车和所述前车的距离；s *为期望距离：
Figure PCTCN2022074581-appb-000056
其中，s 0为跟停最小距离；T为跟车时距；b为舒适减速度。
S410.根据各个目标车辆的期望加速度以及各个目标车辆的风险权重系数计算所述自车的综合加速度。
一种设计中，综合加速度的计算满足以下公式：
a multi＝∑ω i·a i，
其中，i为目标车辆的序号，ω i为目标车辆i的风险权重系数，a i为目标车辆i的期望加速度。
通过输出综合加速度，控制器便可以根据综合加速度来控制自车的行驶状态。由于该综合加速度中考虑了多个目标车辆的期望加速度和风险权重系数，因此可以实现对自车更加准确且安全的控制。
针对L2级别以上的自动驾驶，本申请实施例可输出行驶轨迹以及虚拟目标车辆以便控制器控制自车的行驶状态。请参见图5，图5为本申请实施例提供的又一种智能驾驶方法的流程示意图；步骤S501-步骤S508与图3中的步骤S301-步骤S308相同，此处不再赘述，在步骤S308之后，还包括如下步骤：
S509.输出规划的虚拟行驶场景以及投影到所述虚拟行驶场景的单个虚拟目标车辆的信息。
其中，所述单个虚拟目标车辆的信息根据所述各个目标车辆的风险权重系数和第二行驶状态信息确定；
所述虚拟行驶场景包含规划行驶轨迹，所述规划行驶轨迹是根据所述自车的第一行驶状态信息以及各个目标车辆的第二行驶状态信息规划得到的。
一种设计中，所述虚拟目标车辆的信息包括所述虚拟目标车辆的位置信息、速度信息和加速度信息，所述虚拟目标车辆的位置信息、速度信息和加速度信息，满足以下公式：
(x virtual,y virtual)＝(∑ω i·x i,∑ω i·y i)，
Figure PCTCN2022074581-appb-000057
Figure PCTCN2022074581-appb-000058
其中，i为目标车辆的序号，ω i为目标车辆i的风险权重系数，(x virtual,y virtual)为所述虚拟目标车辆的位置信息，
Figure PCTCN2022074581-appb-000059
为所述虚拟目标车辆的速度信息，
Figure PCTCN2022074581-appb-000060
为所述虚拟目标车辆的加速度信息。
示例性地，可以在车载显示屏上显示虚拟行驶场景，该虚拟行驶场景可模拟实际道路状况，如可包括道路状况信息和各种交通指示牌或指示灯信息等。还包括规划行驶轨迹，而对于预设范围内的多个目标车辆，在虚拟行驶场景中可以输出为单一的一个虚拟目标车辆。实 现更加简洁的显示效果以及更加安全且准确的路径规划和行驶控制。
需要说明的是，在另一些可行的实施例中，图4的步骤S403-步骤S408或图5的步骤S503-步骤S508中对目标车辆的行驶意图预测和风险权重系数调整可以省略，以提高数据处理的效率和精简车辆控制的流程。本申请实施例不作任何限定。
请参照图6，图6为本申请实施例提供的智能驾驶方法与单目标纵向控制的仿真结果对比示意图；其中，横坐标为时间，纵坐标为加速度。标号为1的曲线为常规的单目标纵向控制曲线，标号为2的曲线为根据本申请实施例的方法采用第一套标定参数(包括各种调节系数)生成的纵向控制曲线一，标号为3的曲线为根据本申请实施例的方法采用第二套标定参数(包括各种调节系数)生成的纵向控制曲线二。由图6可知，单目标纵向控制中，遇到需要减速的场景时，减速时间较晚，导致最大加速度的峰值以及变化远远大于本申请实施例的方法，而本申请多目标纵向控制的方法不管是曲线一还是曲线二均会更加提前的进行减速，从而提升了安全性，且最大加速度的峰值以及变化都远小于单目标纵向控制的策略，控制更加的平滑，从而提升了舒适性。因此，可明显看出，本申请实施例的方法在交通拥堵或高速等场景下，可显著提升驾驶的安全性和用户的舒适性。
请参照图7，图7为本申请实施例提供的一种智能驾驶装置的组成示意图；可包括：
获取单元11，用于获取自车的第一行驶状态信息以及多个目标车辆的第二行驶状态信息，所述目标车辆位于预设范围内；
处理单元12，用于根据各个目标车辆的第二行驶状态信息和所述第一行驶状态信息确定各个目标车辆的风险权重系数，所述风险权重系数用于调整所述自车的行驶状态。
在一种可能的实现方式中，所述处理单元12具体用于：
根据所述各个目标车辆对所述自车的影响程度确定各个目标车辆的风险权重系数；
其中，所述影响程度与所述目标车辆的第二行驶状态信息和所述第一行驶状态信息之间满足预先设置的对应关系。
在一种可能的实现方式中，所述第一和/或第二行驶状态信息包含道路状况信息，所述道路状况信息包括以下至少一种：
所述预设范围内的其他车辆信息、路沿信息、车道线信息或可通行障碍物信息。
在一种可能的实现方式中，所述道路状况信息用于表征所述各个目标车辆的行驶意图。
在一种可能的实现方式中，当根据所述道路状况信息确定所述各个目标车辆中的第一目标车辆的行驶意图时，所述处理单元12用于：
分别计算所述道路状况信息中包含的各项信息对所述第一目标车辆的影响程度；
估计所述第一目标车辆向前行驶的意愿程度；
根据所述道路状况信息中包含的各项信息对第一目标车辆的影响程度以及所述第一目标车辆向前行驶的意愿程度，计算所述第一目标车辆受到的综合影响程度；
通过梯度下降方法搜索所述综合影响程度降低的方向；
根据所述综合影响程度降低的方向确定所述第一目标车辆的行驶意图。
在一种可能的实现方式中，所述第一目标车辆的行驶意图通过以下方式影响所述风险权重系数：
若所述第一目标车辆的行驶意图为驶入所述自车当前所处车道，则所述第一目标车辆的风险权重系数提高；或者
若所述第一目标车辆的形式意图为驶出所述自车当前所处车道，则所述第一目标车辆的风险权重系数降低。
在一种可能的实现方式中，所述处理单元12还用于：
根据跟驰模型分别计算各个目标车辆的期望加速度；
根据各个目标车辆的期望加速度以及各个目标车辆的风险权重系数计算所述自车的综合加速度。
在一种可能的实现方式中，所述处理单元12还用于：
输出规划的虚拟行驶场景以及投影到所述虚拟行驶场景的单个虚拟目标车辆的信息，所述单个虚拟目标车辆的信息根据所述各个目标车辆的风险权重系数和第二行驶状态信息确定；
所述虚拟行驶场景包含规划行驶轨迹，所述规划行驶轨迹是根据所述自车的第一行驶状态信息以及各个目标车辆的第二行驶状态信息规划得到的。
在一种可能的实现方式中，所述第一行驶状态信息包括所述自车的速度信息
Figure PCTCN2022074581-appb-000061
和加速度信息
Figure PCTCN2022074581-appb-000062
所述第二行驶状态信息包括目标车辆的位置信息(x i,y i)、速度信息
Figure PCTCN2022074581-appb-000063
和加速度信息
Figure PCTCN2022074581-appb-000064
所述各个目标车辆对所述自车的影响程度，满足以下公式：
Figure PCTCN2022074581-appb-000065
其中，K为欧氏距离且K＝‖x′ i,y′ i‖ 2，x′ i,y′ i为规范化后的位置信息，分别用于表征目标车辆i与所述自车的横向距离和纵向距离，i为目标车辆的序号，系数ω car为用于调节目标车辆的对所述自车的影响程度的调节系数；系数α用于调节目标车辆影响范围大小；
其中，x′ i,y′ i根据横向和纵向的相对速度和加速度计算如下：
Figure PCTCN2022074581-appb-000066
Figure PCTCN2022074581-appb-000067
参数β x,β y,γ x,γ y可分别调节横向速度、纵向速度、横向加速度和纵向加速度对规范化系数∈ x,∈ y的影响程度，∈ x,∈ y越大势能影响范围越小。
在一种可能的实现方式中，当所述道路状况信息中包含路沿信息或车道线信息时，所述处理单元12计算所述路沿线或车道线对所述第一目标车辆的影响程度，满足以下公式：
Figure PCTCN2022074581-appb-000068
其中，系数ω lane为用于调整车道线/路沿线的影响程度的调节系数；δ lane用于调整影响程度衰减的大小；y lane为第一目标车辆与车道线/路沿线的横向距离；
当所述道路状况信息中包含可通行障碍物信息时，所述可通行障碍物对第一目标车辆的影响程度，满足以下公式：
Figure PCTCN2022074581-appb-000069
其中，系数ω obs为调整所述可通行障碍物的影响程度的调节系数；δ x,δ y为分别影响纵向和横向影响程度衰减速度的调节系数；x obs,y obs为所述可通行障碍的物位置。
在一种可能的实现方式中，所述处理单元12估计所述第一目标车辆向前行驶的意愿程度，满足以下公式：
U attr(x)＝-ω attr·x
其中，ω attr为用于调整意愿程度的大小的调节系数。
在一种可能的实现方式中，所述处理单元12根据所述道路状况信息中包含的各项信息对 第一目标车辆的影响程度以及所述第一目标车辆向前行驶的意愿程度，计算所述第一目标车辆受到的综合影响程度，满足以下公式：
Figure PCTCN2022074581-appb-000070
其中，
Figure PCTCN2022074581-appb-000071
表示所述其他车辆对第一目标车辆i的影响程度。
在一种可能的实现方式中，所述处理单元12通过梯度下降方法搜索所述综合影响程度降低的方向，满足以下公式：
Figure PCTCN2022074581-appb-000072
其中，η为梯度下降步长。
在一种可能的实现方式中，所述处理单元12根据跟驰模型分别计算各个目标车辆的期望加速度，满足以下公式：
Figure PCTCN2022074581-appb-000073
其中，v为所述自车的速度；Δv为所述自车和当前车道的前车的速度差；v des为期望速度；σ为标定参数；a max为最大加速度；s为当前时刻所述自车和所述前车的距离；s *为期望距离：
Figure PCTCN2022074581-appb-000074
其中，s 0为跟停最小距离；T为跟车时距；b为舒适减速度。
在一种可能的实现方式中，所述处理单元12根据各个目标车辆的期望加速度以及各个目标车辆的风险权重系数计算所述自车的综合加速度，满足以下公式：
a multi＝∑ω i·a i，
其中，i为目标车辆的序号，ω i为目标车辆i的风险权重系数，a i为目标车辆i的期望加速度。
在一种可能的实现方式中，所述虚拟目标车辆的信息包括所述虚拟目标车辆的位置信息、速度信息和加速度信息，所述虚拟目标车辆的位置信息、速度信息和加速度信息，满足以下公式：
(x virtual,y virtual)＝(∑ω i·x i,∑ω i·y i)，
Figure PCTCN2022074581-appb-000075
Figure PCTCN2022074581-appb-000076
其中，i为目标车辆的序号，ω i为目标车辆i的风险权重系数，(x virtual,y virtual)为所述虚拟目标车辆的位置信息，
Figure PCTCN2022074581-appb-000077
为所述虚拟目标车辆的速度信息，
Figure PCTCN2022074581-appb-000078
为所述虚拟目标车辆的加速度信息。
在一种可能的实现方式中，各种调节系数的标定根据来自外部的输入参数确定。
在一种可能的实现方式中，各种调节系数的标定根据如下目标函数确定：
Figure PCTCN2022074581-appb-000079
所述目标函数用于确定t时刻对于自车实际驾驶方向的最小化影响程度下降方向，其中，在t时刻所述自车方向向量为
Figure PCTCN2022074581-appb-000080
对所述自车影响程度梯度下降方向
Figure PCTCN2022074581-appb-000081
该智能驾驶装置所涉及的与本申请实施例提供的技术方案相关的概念，解释和详细说明及其他步骤请参见前述方法实施例中关于这些内容的描述，此处不做赘述。
请参见图8，为本申请实施例提供的另一种智能驾驶装置的组成示意图；
该智能驾驶装置可以包括至少一个处理器110以及通信接口120。所述通信接口120用于提供信息的输入或输出，所述至少一个处理器110用于执行程序或指令，以使得所述处理器110执行如上图2-图5对应的方法中智能驾驶装置所在的设备执行的方法。
例如，通信接口120可以获取各类传感器或摄像头采集的数据如自车的第一行驶状态信息以及预设范围内多个目标车辆的第二状态信息，还可以输出控制数据如综合加速度或规划行驶轨迹给控制器。
进一步地，所述装置还可以包括存储器，存储器可用于存储处理器110执行的指令，进一步地还可以用于存储通信接口120获取到的数据以及处理器110处理得到的数据。
进一步地，所述装置还可以包括总线，所述装置中的各个部件可以通过总线连接。
可选地，所述装置可以是一个芯片，包括一个或多个处理电路。所述一个或多个处理电路用于实现如图2-图5中任一项所述的方法。所述通信接口120可以为芯片的输入/输出电路或者接口。
所述装置存在的设备包括上述芯片，该设备可以是中央控制单元等，也可以是车辆上用于实现本申请实施例中智能驾驶方法独立设置的设备，本申请实施例不作任何限定。
作为一种实现方式，通信接口120的功能可以考虑通过收发电路或者收发的专用芯片实现。处理器110可以考虑通过专用处理芯片、处理电路、处理器或者通用芯片实现。
作为另一种实现方式，可以考虑使用通用计算机的方式来实现本申请实施例提供的智能驾驶装置。即将实现处理器110，通信接口120功能的程序代码存储在存储器中，通用处理器通过执行存储器中的代码来实现处理器110和通信接口120的功能。
该智能驾驶装置所涉及的与本申请实施例提供的技术方案相关的概念，解释和详细说明及其他步骤请参见前述方法或其他实施例中关于这些内容的描述，此处不做赘述。
本领域技术人员可以理解，为了便于说明，图8仅示出了一个处理器。在实际的装置中，可以存在多个处理器和存储器。存储器也可以称为存储介质或者存储设备等，本申请实施例对此不做限制。在本申请实施例中，处理器可以是中央处理单元(Central Processing Unit，简称为“CPU”)，该处理器还可以是其他通用处理器、数字信号处理器(DSP)、专用集成电路(ASIC)、现成可编程门阵列(FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件等。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。该存储器可以包括只读存储器和随机存取存储器，并向处理器提供指令和数据。存储器的一部分还可以包括非易失性随机存取存储器。该总线除包括数据总线之外，还可以包括电源总线、控制总线和状态信号总线等。但是为了清楚说明起见，在图中将各种总线都标为总线。
根据本申请实施例提供的方法、智能驾驶装置和控制器，本申请实施例还提供一种车辆的控制系统，各个设备之间的关系和指令流程可以参见图1-图5实施例的描述和说明，此处不再赘述。
此外，根据本申请实施例提供的方法、智能驾驶装置和控制器，本申请实施例还提供一种车辆，车辆中包括的各个设备之间的关系和指令流程可以参见图1-图5实施例的描述和说明，此处不再赘述。
应理解，在本申请的各种实施例中，上述各过程的序号的大小并不意味着执行顺序的先后，各过程的执行顺序应以其功能和内在逻辑确定，而不应对本申请实施例的实施过程构成 任何限定。
本领域普通技术人员可以意识到，结合本文中所公开的实施例描述的各种说明性逻辑块(illustrative logical block)和步骤(step)，能够以电子硬件、或者计算机软件和电子硬件的结合来实现。这些功能究竟以硬件还是软件方式来执行，取决于技术方案的特定应用和设计约束条件。专业技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能，但是这种实现不应认为超出本申请的范围。
在本申请所提供的几个实施例中，应该理解到，所揭露的系统、装置和方法，可以通过其它的方式实现。例如，以上所描述的装置实施例仅仅是示意性的，例如，所述单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个系统，或一些特征可以忽略，或不执行。另一点，所显示或讨论的相互之间的耦合或直接耦合或通信连接可以是通过一些接口，装置或单元的间接耦合或通信连接，可以是电性，机械或其它的形式。
在上述实施例中，可以全部或部分地通过软件、硬件、固件或者其任意组合来实现。当使用软件实现时，可以全部或部分地以计算机程序产品的形式实现。所述计算机程序产品包括一个或多个计算机指令。在计算机上加载和执行所述计算机程序指令时，全部或部分地产生按照本申请实施例所述的流程或功能。所述计算机可以是通用计算机、专用计算机、计算机网络、或者其他可编程装置。所述计算机指令可以存储在计算机可读存储介质中，或者从一个计算机可读存储介质向另一个计算机可读存储介质传输，例如，所述计算机指令可以从一个网站站点、计算机、服务器或数据中心通过有线(例如同轴电缆、光纤、数字用户线(DSL))或无线(例如红外、无线、微波等)方式向另一个网站站点、计算机、服务器或数据中心进行传输。所述计算机可读存储介质可以是计算机能够存取的任何可用介质或者是包含一个或多个可用介质集成的服务器、数据中心等数据存储设备。所述可用介质可以是磁性介质，(例如，软盘、硬盘、磁带)、光介质(例如，DVD)、或者半导体介质(例如固态硬盘Solid State Disk(SSD))等。
以上所述，仅为本申请的具体实施方式，但本申请的保护范围并不局限于此，任何熟悉本技术领域的技术人员在本申请揭露的技术范围内，可轻易想到变化或替换，都应涵盖在本申请的保护范围之内。因此，本申请的保护范围应以所述权利要求的保护范围为准。

Claims (30)

  1. 一种智能驾驶方法，其特征在于，包括：
    获取自车的第一行驶状态信息以及多个目标车辆的第二行驶状态信息，所述目标车辆位于预设范围内；
    根据各个目标车辆的第二行驶状态信息和所述第一行驶状态信息确定各个目标车辆的风险权重系数，所述风险权重系数用于调整所述自车的行驶状态。
  2. 根据权利要求1所述的方法，其特征在于，所述根据各个目标车辆的第二行驶状态信息和所述第一行驶状态信息确定各个目标车辆的风险权重系数，包括：
    根据所述各个目标车辆对所述自车的影响程度确定各个目标车辆的风险权重系数；
    其中，所述影响程度与所述目标车辆的第二行驶状态信息和所述第一行驶状态信息之间满足预先设置的对应关系。
  3. 根据权利要求1或2所述的方法，其特征在于，所述第一和/或第二行驶状态信息包含道路状况信息，所述道路状况信息包括以下至少一种：
    所述预设范围内的其他车辆信息、路沿信息、车道线信息或可通行障碍物信息。
  4. 根据权利要求3所述的方法，其特征在于，所述道路状况信息用于表征所述各个目标车辆的行驶意图。
  5. 根据权利要求3或4所述的方法，其特征在于，所述方法还包括：
    分别计算所述道路状况信息中包含的各项信息对第一目标车辆的影响程度；
    估计所述第一目标车辆向前行驶的意愿程度；
    根据所述道路状况信息中包含的各项信息对第一目标车辆的影响程度以及所述第一目标车辆向前行驶的意愿程度，计算所述第一目标车辆受到的综合影响程度；
    通过梯度下降方法搜索所述综合影响程度降低的方向；
    根据所述综合影响程度降低的方向确定所述第一目标车辆的行驶意图。
  6. 根据权利要求5所述的方法，其特征在于，所述第一目标车辆的行驶意图通过以下方式影响所述风险权重系数：
    若所述第一目标车辆的行驶意图为驶入所述自车当前所处车道，则所述第一目标车辆的风险权重系数提高；或者
    若所述第一目标车辆的形式意图为驶出所述自车当前所处车道，则所述第一目标车辆的风险权重系数降低。
  7. 根据权利要求1-6任一项所述的方法，其特征在于，所述方法还包括：
    根据跟驰模型分别计算各个目标车辆的期望加速度；
    根据各个目标车辆的期望加速度以及各个目标车辆的风险权重系数计算所述自车的综合加速度。
  8. 根据权利要求1-6任一项所述的方法，其特征在于，所述方法还包括：
    输出规划的虚拟行驶场景以及投影到所述虚拟行驶场景的单个虚拟目标车辆的信息，所述单个虚拟目标车辆的信息根据所述各个目标车辆的风险权重系数和第二行驶状态信息确定；
    所述虚拟行驶场景包含规划行驶轨迹，所述规划行驶轨迹是根据所述自车的第一行驶状态信息以及各个目标车辆的第二行驶状态信息规划得到的。
  9. 根据权利要求2-8任一项所述的方法，其特征在于，所述第一行驶状态信息包括所述自车的速度信息
    Figure PCTCN2022074581-appb-100001
    和加速度信息
    Figure PCTCN2022074581-appb-100002
    所述第二行驶状态信息包括目标车辆的位置信息
    Figure PCTCN2022074581-appb-100003
    速度信息
    Figure PCTCN2022074581-appb-100004
    和加速度信息
    Figure PCTCN2022074581-appb-100005
    所述各个目标车辆对所述自车的影响程度，满足以下公式：
    Figure PCTCN2022074581-appb-100006
    其中，K为欧氏距离且K＝‖x ′i,y′ i‖ 2，x′ i,y′ i为规范化后的位置信息，分别用于表征目标车辆i与所述自车的横向距离和纵向距离，i为目标车辆的序号，系数ω car为用于调节目标车辆的对所述自车的影响程度的调节系数；系数α用于调节目标车辆影响范围大小；
    其中，x′ i,y′ i根据横向和纵向的相对速度和加速度计算如下：
    Figure PCTCN2022074581-appb-100007
    Figure PCTCN2022074581-appb-100008
    参数β x,β y,γ x,γ y可分别调节横向速度、纵向速度、横向加速度和纵向加速度对规范化系数∈ x，∈ y的影响程度，∈ x,∈ y越大势能影响范围越小。
  10. 根据权利要求5所述的方法，其特征在于，所述分别计算所述道路状况信息中包含的各项信息对第一目标车辆的影响程度，包括：
    当所述道路状况信息中包含路沿信息或车道线信息时，所述路沿线或车道线对所述第一目标车辆的影响程度，满足以下公式：
    Figure PCTCN2022074581-appb-100009
    其中，系数ω lane为用于调整车道线/路沿线的影响程度的调节系数；δ lane用于调整影响程度衰减的大小；y lane为第一目标车辆与车道线/路沿线的横向距离；
    当所述道路状况信息中包含可通行障碍物信息时，所述可通行障碍物对第一目标车辆的影响程度，满足以下公式：
    Figure PCTCN2022074581-appb-100010
    其中，系数ω obs为调整所述可通行障碍物的影响程度的调节系数；δ x,δ y为分别影响纵向和横向影响程度衰减速度的调节系数；x obs,y obs为所述可通行障碍的物位置。
  11. 根据权利要求5或10所述的方法，其特征在于，所述估计所述第一目标车辆向前行驶的意愿程度，满足以下公式：
    U attr(x)＝-ω attr·x
    其中，ω attr为用于调整意愿程度的大小的调节系数。
  12. 根据权利要求10或11所述的方法，其特征在于，所述根据所述道路状况信息中包含的各项信息对第一目标车辆的影响程度以及所述第一目标车辆向前行驶的意愿程度，计算所述第一目标车辆受到的综合影响程度，满足以下公式：
    Figure PCTCN2022074581-appb-100011
    其中，
    Figure PCTCN2022074581-appb-100012
    表示所述其他车辆对第一目标车辆i的影响程度。
  13. 根据权利要求12所述的方法，其特征在于，所述通过梯度下降方法搜索所述综合影响程度降低的方向，满足以下公式：
    Figure PCTCN2022074581-appb-100013
    其中，η为梯度下降步长。
  14. 根据权利要求7所述的方法，其特征在于，所述根据跟驰模型分别计算各个目标车辆的期望加速度，满足以下公式：
    Figure PCTCN2022074581-appb-100014
    其中，v为所述自车的速度；Δv为所述自车和当前车道的前车的速度差；v des为期望速度；σ为标定参数；a max为最大加速度；s为当前时刻所述自车和所述前车的距离；s *为期望距离：
    Figure PCTCN2022074581-appb-100015
    其中，s 0为跟停最小距离；T为跟车时距；b为舒适减速度。
  15. 根据权利要求14所述的方法，其特征在于，所述根据各个目标车辆的期望加速度以
    及各个目标车辆的风险权重系数计算所述自车的综合加速度，满足以下公式：
    a multi＝∑ω i·a i，
    其中，i为目标车辆的序号，ω i为目标车辆i的风险权重系数，a i为目标车辆i的期望加速度。
  16. 根据权利要求8所述的方法，其特征在于，所述虚拟目标车辆的信息包括所述虚拟目标车辆的位置信息、速度信息和加速度信息，所述虚拟目标车辆的位置信息、速度信息和加速度信息，满足以下公式：
    (x virtual,y virtual)＝(∑ω i·x i,∑ω i·y i)，
    Figure PCTCN2022074581-appb-100016
    Figure PCTCN2022074581-appb-100017
    其中，i为目标车辆的序号，ω i为目标车辆i的风险权重系数，(x virtual,y virtual)为所述虚拟目标车辆的位置信息，
    Figure PCTCN2022074581-appb-100018
    为所述虚拟目标车辆的速度信息，
    Figure PCTCN2022074581-appb-100019
    为所述虚拟目标车辆的加速度信息。
  17. 根据权利要求9、10、11或12所述的方法，其特征在于，各种调节系数的标定根据来自外部的输入参数确定。
  18. 根据权利要求9、10、11或12所述的方法，其特征在于，各种调节系数的标定根据如下目标函数确定：
    Figure PCTCN2022074581-appb-100020
    所述目标函数用于确定t时刻对于自车实际驾驶方向的最小化影响程度下降方向，其中，在t时刻所述自车方向向量为
    Figure PCTCN2022074581-appb-100021
    对所述自车影响程度梯度下降方向
    Figure PCTCN2022074581-appb-100022
  19. 一种智能驾驶装置，其特征在于，包括：
    获取单元，用于获取自车的第一行驶状态信息以及多个目标车辆的第二行驶状态信息，所述目标车辆位于预设范围内；
    处理单元，用于根据各个目标车辆的第二行驶状态信息和所述第一行驶状态信息确定各个目标车辆的风险权重系数，所述风险权重系数用于调整所述自车的行驶状态。
  20. 根据权利要求19所述的装置，其特征在于，所述处理单元具体用于：
    根据所述各个目标车辆对所述自车的影响程度确定各个目标车辆的风险权重系数；
    其中，所述影响程度与所述目标车辆的第二行驶状态信息和所述第一行驶状态信息之间满足预先设置的对应关系。
  21. 根据权利要求19或20所述的装置，其特征在于，所述第一和/或第二行驶状态信息包含道路状况信息，所述道路状况信息包括以下至少一种：
    所述预设范围内的其他车辆信息、路沿信息、车道线信息或可通行障碍物信息。
  22. 根据权利要求21所述的装置，其特征在于，所述道路状况信息用于表征所述各个目标车辆的行驶意图。
  23. 根据权利要求22所述的装置，其特征在于，当根据所述道路状况信息确定所述各个目标车辆中的第一目标车辆的行驶意图时，所述处理单元用于：
    分别计算所述道路状况信息中包含的各项信息对所述第一目标车辆的影响程度；
    估计所述第一目标车辆向前行驶的意愿程度；
    根据所述道路状况信息中包含的各项信息对第一目标车辆的影响程度以及所述第一目标车辆向前行驶的意愿程度，计算所述第一目标车辆受到的综合影响程度；
    通过梯度下降方法搜索所述综合影响程度降低的方向；
    根据所述综合影响程度降低的方向确定所述第一目标车辆的行驶意图。
  24. 根据权利要求23所述的装置，其特征在于，所述第一目标车辆的行驶意图通过以下方式影响所述风险权重系数：
    若所述第一目标车辆的行驶意图为驶入所述自车当前所处车道，则所述第一目标车辆的风险权重系数提高；或者
    若所述第一目标车辆的形式意图为驶出所述自车当前所处车道，则所述第一目标车辆的风险权重系数降低。
  25. 根据权利要求19-24任一项所述的装置，其特征在于，所述处理单元还用于：
    根据跟驰模型分别计算各个目标车辆的期望加速度；
    根据各个目标车辆的期望加速度以及各个目标车辆的风险权重系数计算所述自车的综合加速度。
  26. 根据权利要求19-24任一项所述的装置，其特征在于，所述处理单元还用于：
    输出规划的虚拟行驶场景以及投影到所述虚拟行驶场景的单个虚拟目标车辆的信息，所述单个虚拟目标车辆的信息根据所述各个目标车辆的风险权重系数和第二行驶状态信息确定；
    所述虚拟行驶场景包含规划行驶轨迹，所述规划行驶轨迹是根据所述自车的第一行驶状态信息以及各个目标车辆的第二行驶状态信息规划得到的。
  27. 一种智能驾驶装置，其特征在于，包括：
    至少一个处理器以及通信接口，所述通信接口用于提供信息的输入或输出，所述至少一个处理器用于执行程序或指令，以使得所述装置所在的设备执行如权利要求1-18任一项所述的方法。
  28. 一种计算机可读存储介质，其特征在于，包括：
    所述计算机可读存储介质中存储有指令，当其在计算机上运行时，实现如权利要求1-18任一项所述的方法。
  29. 一种芯片，包括一个或多个处理电路，其中，所述一个或多个处理电路用于实现如权利要求1-18中任一项所述的方法。
  30. 一种车辆，其特征在于，包括：
    如权利要求19-26任一项所述的装置。
PCT/CN2022/074581 2022-01-28 2022-01-28 一种智能驾驶方法、装置及车辆 WO2023141940A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/CN2022/074581 WO2023141940A1 (zh) 2022-01-28 2022-01-28 一种智能驾驶方法、装置及车辆

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2022/074581 WO2023141940A1 (zh) 2022-01-28 2022-01-28 一种智能驾驶方法、装置及车辆

Publications (1)

Publication Number Publication Date
WO2023141940A1 true WO2023141940A1 (zh) 2023-08-03

Family

ID=87469961

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/074581 WO2023141940A1 (zh) 2022-01-28 2022-01-28 一种智能驾驶方法、装置及车辆

Country Status (1)

Country Link
WO (1) WO2023141940A1 (zh)

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108437983A (zh) * 2018-03-29 2018-08-24 吉林大学 一种基于预测安全的智能车辆避障系统
CN110239550A (zh) * 2019-05-28 2019-09-17 浙江吉利控股集团有限公司 一种车辆自动变道方法
CN110422168A (zh) * 2019-08-08 2019-11-08 智邮开源通信研究院(北京)有限公司 车道识别系统、方法及自动驾驶汽车
US20200139959A1 (en) * 2018-11-02 2020-05-07 Zoox, Inc. Cost scaling in trajectory generation
US20200172106A1 (en) * 2018-12-04 2020-06-04 GM Global Technology Operations LLC System and method for control of an autonomous vehicle
CN111830951A (zh) * 2019-03-29 2020-10-27 中科院微电子研究所昆山分所 一种自适应跟随预测控制方法、系统及装置
CN114013443A (zh) * 2021-11-12 2022-02-08 哈尔滨工业大学 一种基于分层强化学习的自动驾驶车辆换道决策控制方法

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108437983A (zh) * 2018-03-29 2018-08-24 吉林大学 一种基于预测安全的智能车辆避障系统
US20200139959A1 (en) * 2018-11-02 2020-05-07 Zoox, Inc. Cost scaling in trajectory generation
US20200172106A1 (en) * 2018-12-04 2020-06-04 GM Global Technology Operations LLC System and method for control of an autonomous vehicle
CN111830951A (zh) * 2019-03-29 2020-10-27 中科院微电子研究所昆山分所 一种自适应跟随预测控制方法、系统及装置
CN110239550A (zh) * 2019-05-28 2019-09-17 浙江吉利控股集团有限公司 一种车辆自动变道方法
CN110422168A (zh) * 2019-08-08 2019-11-08 智邮开源通信研究院(北京)有限公司 车道识别系统、方法及自动驾驶汽车
CN114013443A (zh) * 2021-11-12 2022-02-08 哈尔滨工业大学 一种基于分层强化学习的自动驾驶车辆换道决策控制方法

Similar Documents

Publication Publication Date Title
US11941873B2 (en) Determining drivable free-space for autonomous vehicles
US11204607B2 (en) Trajectory setting device and trajectory setting method
US11604470B2 (en) Safety procedure analysis for obstacle avoidance in autonomous vehicles
WO2021008605A1 (zh) 一种确定车速的方法和装置
WO2019157193A1 (en) Controlling autonomous vehicles using safe arrival times
WO2022183329A1 (zh) 一种智能驾驶方法、装置、存储介质及计算机程序
CN112512887B (zh) 一种行驶决策选择方法以及装置
WO2022205243A1 (zh) 一种变道区域获取方法以及装置
JP7207256B2 (ja) 車両制御システム
JP2022180294A (ja) 自律型車両のためのモーション・プランニング軌道において到着時間及び安全性手順を使用すること
US11753035B2 (en) Vehicle control system
WO2022052856A1 (zh) 基于交通工具的数据处理方法、装置、计算机及存储介质
JP7379033B2 (ja) 運転支援方法及び運転支援装置
WO2021233154A1 (zh) 可行驶区域检测方法、装置、设备及存储介质
WO2023141940A1 (zh) 一种智能驾驶方法、装置及车辆
JP7147448B2 (ja) 地図情報システム
CN115431961A (zh) 车辆的控制方法、装置、车辆及存储介质
WO2021170140A1 (zh) 一种车道结构检测方法及装置
TWI823613B (zh) 防撞預警方法、車載裝置及儲存介質
US20240025444A1 (en) Path generation based on predicted actions
US20240025395A1 (en) Path generation based on predicted actions
US20240025443A1 (en) Path generation based on predicted actions
WO2023015505A1 (zh) 一种车辆驾驶方法、装置及系统
JP2019214291A (ja) 走行支援方法及び走行支援装置
WO2023053165A1 (ja) 運転支援装置及びコンピュータプログラムを記録した記録媒体

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: 22922784

Country of ref document: EP

Kind code of ref document: A1