WO2019174397A1 - 自动驾驶安全评估方法、装置和系统 - Google Patents

自动驾驶安全评估方法、装置和系统 Download PDF

Info

Publication number
WO2019174397A1
WO2019174397A1 PCT/CN2019/072074 CN2019072074W WO2019174397A1 WO 2019174397 A1 WO2019174397 A1 WO 2019174397A1 CN 2019072074 W CN2019072074 W CN 2019072074W WO 2019174397 A1 WO2019174397 A1 WO 2019174397A1
Authority
WO
WIPO (PCT)
Prior art keywords
vehicle
risk
time
driving mode
automatic driving
Prior art date
Application number
PCT/CN2019/072074
Other languages
English (en)
French (fr)
Inventor
佘晓丽
蔡建永
沈骏强
陈奇
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 EP19768393.1A priority Critical patent/EP3745378B1/en
Priority to KR1020207026721A priority patent/KR102574105B1/ko
Priority to JP2020549588A priority patent/JP7227267B2/ja
Priority to BR112020017519-6A priority patent/BR112020017519A2/pt
Publication of WO2019174397A1 publication Critical patent/WO2019174397A1/zh
Priority to US17/018,505 priority patent/US11872999B2/en

Links

Images

Classifications

    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W50/00Details 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
    • B60W50/04Monitoring the functioning of the control system
    • B60W50/045Monitoring control system parameters
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W60/00Drive control systems specially adapted for autonomous road vehicles
    • B60W60/005Handover processes
    • B60W60/0059Estimation of the risk associated with autonomous or manual driving, e.g. situation too complex, sensor failure or driver incapacity
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W30/00Purposes 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
    • B60W30/08Active safety systems predicting or avoiding probable or impending collision or attempting to minimise its consequences
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W30/00Purposes 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
    • B60W30/08Active safety systems predicting or avoiding probable or impending collision or attempting to minimise its consequences
    • B60W30/095Predicting travel path or likelihood of collision
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W30/00Purposes 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
    • B60W30/18Propelling the vehicle
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W30/00Purposes 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
    • B60W30/18Propelling the vehicle
    • B60W30/18009Propelling the vehicle related to particular drive situations
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W30/00Purposes 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
    • B60W30/18Propelling the vehicle
    • B60W30/18009Propelling the vehicle related to particular drive situations
    • B60W30/18145Cornering
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W30/00Purposes 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
    • B60W30/18Propelling the vehicle
    • B60W30/18009Propelling the vehicle related to particular drive situations
    • B60W30/18159Traversing an intersection
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W30/00Purposes 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
    • B60W30/18Propelling the vehicle
    • B60W30/18009Propelling the vehicle related to particular drive situations
    • B60W30/18163Lane change; Overtaking manoeuvres
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W30/00Purposes 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
    • B60W30/18Propelling the vehicle
    • B60W30/182Selecting between different operative modes, e.g. comfort and performance modes
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W50/00Details 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
    • B60W50/08Interaction between the driver and the control system
    • B60W50/082Selecting or switching between different modes of propelling
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W50/00Details 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
    • B60W50/08Interaction between the driver and the control system
    • B60W50/10Interpretation of driver requests or demands
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W60/00Drive control systems specially adapted for autonomous road vehicles
    • B60W60/001Planning or execution of driving tasks
    • B60W60/0011Planning or execution of driving tasks involving control alternatives for a single driving scenario, e.g. planning several paths to avoid obstacles
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00Error detection; Error correction; Monitoring
    • G06F11/30Monitoring
    • G06F11/3003Monitoring arrangements specially adapted to the computing system or computing system component being monitored
    • G06F11/3013Monitoring arrangements specially adapted to the computing system or computing system component being monitored where the computing system is an embedded system, i.e. a combination of hardware and software dedicated to perform a certain function in mobile devices, printers, automotive or aircraft systems
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION 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
    • G06Q10/00Administration; Management
    • G06Q10/04Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION 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
    • G06Q10/00Administration; Management
    • G06Q10/04Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G06Q10/047Optimisation of routes or paths, e.g. travelling salesman problem
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION 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
    • G06Q10/00Administration; Management
    • G06Q10/06Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION 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
    • G06Q10/00Administration; Management
    • G06Q10/06Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063Operations research, analysis or management
    • G06Q10/0635Risk analysis of enterprise or organisation activities
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION 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
    • G06Q50/00Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/10Services
    • G06Q50/26Government or public services
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION 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
    • G06Q50/00Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/10Services
    • G06Q50/26Government or public services
    • G06Q50/265Personal security, identity or safety
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION 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
    • G06Q50/00Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/40Business processes related to the transportation industry
    • GPHYSICS
    • G08SIGNALLING
    • G08GTRAFFIC CONTROL SYSTEMS
    • G08G1/00Traffic control systems for road vehicles
    • G08G1/16Anti-collision systems
    • GPHYSICS
    • G08SIGNALLING
    • G08GTRAFFIC CONTROL SYSTEMS
    • G08G1/00Traffic control systems for road vehicles
    • G08G1/16Anti-collision systems
    • G08G1/166Anti-collision systems for active traffic, e.g. moving vehicles, pedestrians, bikes
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W2520/00Input parameters relating to overall vehicle dynamics
    • B60W2520/10Longitudinal speed
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W2520/00Input parameters relating to overall vehicle dynamics
    • B60W2520/10Longitudinal speed
    • B60W2520/105Longitudinal acceleration
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W2520/00Input parameters relating to overall vehicle dynamics
    • B60W2520/12Lateral speed
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W2520/00Input parameters relating to overall vehicle dynamics
    • B60W2520/12Lateral speed
    • B60W2520/125Lateral acceleration
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W2552/00Input parameters relating to infrastructure
    • B60W2552/05Type of road, e.g. motorways, local streets, paved or unpaved roads
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W2552/00Input parameters relating to infrastructure
    • B60W2552/30Road curve radius
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W2554/00Input parameters relating to objects
    • B60W2554/40Dynamic objects, e.g. animals, windblown objects
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W2554/00Input parameters relating to objects
    • B60W2554/40Dynamic objects, e.g. animals, windblown objects
    • B60W2554/402Type
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W2554/00Input parameters relating to objects
    • B60W2554/80Spatial relation or speed relative to objects
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W2554/00Input parameters relating to objects
    • B60W2554/80Spatial relation or speed relative to objects
    • B60W2554/802Longitudinal distance
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W2554/00Input parameters relating to objects
    • B60W2554/80Spatial relation or speed relative to objects
    • B60W2554/803Relative lateral speed
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W2554/00Input parameters relating to objects
    • B60W2554/80Spatial relation or speed relative to objects
    • B60W2554/804Relative longitudinal speed
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W2555/00Input parameters relating to exterior conditions, not covered by groups B60W2552/00, B60W2554/00
    • B60W2555/20Ambient conditions, e.g. wind or rain
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W2555/00Input parameters relating to exterior conditions, not covered by groups B60W2552/00, B60W2554/00
    • B60W2555/60Traffic rules, e.g. speed limits or right of way
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W2556/00Input parameters relating to data
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W2556/00Input parameters relating to data
    • B60W2556/10Historical data
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W2556/00Input parameters relating to data
    • B60W2556/45External transmission of data to or from the vehicle
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT 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
    • B60W50/00Details 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
    • B60W50/08Interaction between the driver and the control system
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00Error detection; Error correction; Monitoring
    • G06F11/07Responding to the occurrence of a fault, e.g. fault tolerance
    • G06F11/0703Error or fault processing not based on redundancy, i.e. by taking additional measures to deal with the error or fault not making use of redundancy in operation, in hardware, or in data representation
    • G06F11/0706Error or fault processing not based on redundancy, i.e. by taking additional measures to deal with the error or fault not making use of redundancy in operation, in hardware, or in data representation the processing taking place on a specific hardware platform or in a specific software environment
    • G06F11/0736Error or fault processing not based on redundancy, i.e. by taking additional measures to deal with the error or fault not making use of redundancy in operation, in hardware, or in data representation the processing taking place on a specific hardware platform or in a specific software environment in functional embedded systems, i.e. in a data processing system designed as a combination of hardware and software dedicated to performing a certain function
    • G06F11/0739Error or fault processing not based on redundancy, i.e. by taking additional measures to deal with the error or fault not making use of redundancy in operation, in hardware, or in data representation the processing taking place on a specific hardware platform or in a specific software environment in functional embedded systems, i.e. in a data processing system designed as a combination of hardware and software dedicated to performing a certain function in a data processing system embedded in automotive or aircraft systems
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00Error detection; Error correction; Monitoring
    • G06F11/07Responding to the occurrence of a fault, e.g. fault tolerance
    • G06F11/0703Error or fault processing not based on redundancy, i.e. by taking additional measures to deal with the error or fault not making use of redundancy in operation, in hardware, or in data representation
    • G06F11/0751Error or fault detection not based on redundancy
    • G06F11/0754Error or fault detection not based on redundancy by exceeding limits
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00Error detection; Error correction; Monitoring
    • G06F11/36Preventing errors by testing or debugging software
    • G06F11/3668Software testing
    • G06F11/3672Test management
    • G06F11/3692Test management for test results analysis
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F2201/00Indexing scheme relating to error detection, to error correction, and to monitoring
    • G06F2201/81Threshold

Definitions

  • the present application relates to the field of automatic driving, and in particular, to an automatic driving safety assessment method, apparatus and system.
  • the self-driving car also known as the driverless car, is capable of acquiring environmental information through sensors such as radar, and controlling the driving of the vehicle according to the environmental information, thereby reducing the time for the driver to control the car.
  • Self-driving cars have the advantage of reducing traffic violations.
  • the driver controls the car in the automatic driving mode for less time than the driver controls the car in the traditional manual driving mode there are still many uncertainties in the self-driving car, in order to reduce The risk of passengers using autonomous vehicles requires a rigorous assessment of their safety before selling autonomous vehicles.
  • An existing automatic driving safety evaluation method is to input a manual driving route into a simulated automatic driving system, and respectively calculate safety-related factors in the artificial driving mode and the automatic driving mode, and obtain quantitative values of the artificial driving mode and the automatic driving mode.
  • the above method can use the data in the manual driving mode to evaluate the safety of the automatic driving mode, and reduce the cycle time of the mass production of the self-driving car.
  • the safety evaluation based on the simulated automatic driving system cannot restore the real driver control scene. As a result, the safety evaluation results obtained based on the simulated automatic driving system have a large deviation from the true automatic driving safety.
  • the present application provides a method and apparatus for automatic driving safety assessment that can reduce the deviation of safety assessment results from true autonomous driving safety.
  • an automatic driving safety assessment method comprising: obtaining r b and r c , wherein r b is a risk value of a shadow driving mode of a vehicle in a first unit of measurement, and r c is within a first unit of measurement
  • the vehicle is based on the risk value of the automatic driving mode of the preset route, and the shadow driving mode is an automatic driving mode of the vehicle based on the real-time route, the real-time route is a route predicted according to the position and the motion parameter of the vehicle in the manual driving mode, the first measurement
  • the unit is a time period or a distance; determining R B according to r b , R B is the risk value of the shadow driving mode of the vehicle in multiple measuring units, and the plurality of measuring units includes the first unit of measurement; determining R C , R according to r c C is a risk value of the preset driving-based automatic driving mode of the vehicle in a plurality of measuring units; wherein R B and R C are used to determine whether
  • the automatic driving safety assessment method provided by the present application collects the risk data generated by the automatic driving mode based on the preset route, and also collects the automatic driving mode based on the real-time route.
  • Risk data the real-time route is a route predicted according to the position and motion parameters of the vehicle in the manual driving mode. For example, when the driver controls the car to change the preset route, the vehicle safety evaluation device can re-plan the automatic driving route according to the driver's control route. And determine the safety of the automatic driving based on the re-planned automatic driving route, so that the safety of the automatic driving can be more realistically evaluated.
  • the method further includes: calculating a sum of R B and R C ; determining that the safety of the vehicle's automatic driving mode is consistent when the sum of R B and R C is less than or equal to an automatic driving risk threshold Claim.
  • the above embodiment can be implemented by means for determining R B and R C without transmitting R B and R C to other devices and evaluating the safety of the vehicle's automatic driving mode by other devices, thereby improving the real-time safety assessment of the automatic driving mode. Sex.
  • the automatic driving risk threshold is a(R A +R′ A ), where a is a risk tolerance coefficient, and R A is a risk of a manual driving mode of the vehicle in multiple units of measurement.
  • the value is that the road section that the vehicle passes in the plurality of measurement units is the first road section, and R' A is the risk value of the vehicle passing the first road section based on the manual driving mode outside the plurality of measurement units.
  • the risk value of the automatic driving can be compared with the risk value of the manual driving, and the tolerance coefficient of the automatic driving risk value can be accepted, so that the safety of the automatic driving can be judged whether the safety is satisfied.
  • determining R B according to r b and determining R C according to r c include: Determining R B or R C , where event represents a risk scenario, s event represents a coefficient of severity of the accident in the risk scenario, X represents the plurality of units of measure, and when r i represents r b , R i represents R B When r i represents r c , R i represents R C .
  • the s event reflects the severity of the accident in different risk scenarios. For example, in a collision scene, the severity of the collision is much higher than the severity of the collision with other objects. Therefore, the severity of the collision can be set to 1, and the collision The severity of other objects can be selected, for example, from 0.1-0.9 depending on the nature of the object; r i reflects the probability of the risk being converted into an accident, ⁇ reflects the cumulative risk of multiple units of measure, and R i reflects the average The probability that a risk is converted into an accident within the unit of measure.
  • the above formula can objectively reflect the risk of autonomous driving.
  • obtaining r b and r c includes: determining r b and r c according to at least one of the following parameters when the risk scene is a collision scenario: a speed ⁇ v of the vehicle relative to the collision object, based on the collision
  • the trajectory of the object is estimated by the time to collision (TTC) of the collision of the vehicle with the collision object, the time to headway (THW) of the vehicle based on the distance between the vehicle and the collision object, and the distance between the vehicle and the colliding object, wherein, Delta] v and r b r c or positive correlation, the TTC or negative correlation with r b r c, THW or negative correlation with r b r c, the distance between the vehicle and the colliding object and r b or r c is negatively correlated.
  • TTC time to collision occurrence
  • THW time to collision occurrence
  • vehicle and collision object The farther the distance is, the more likely the vehicle adjusts the travel route to avoid collisions, and the less likely the collision occurs, so the distance between the vehicle and the collision object is negatively correlated with r b or r c .
  • r b and r c are determined according to at least one of the following parameters: speed ⁇ v of the vehicle relative to the collision object, time required for the collision of the vehicle and the collision object based on the motion trajectory of the collision object
  • obtaining r b and r c includes: when the risk scene is a traffic signal scene, determining r b and r c according to the speed of the vehicle and the distance of the vehicle to the traffic signal, wherein the speed of the vehicle is r b or r c is positively correlated, and the distance from the vehicle to the traffic signal is negatively correlated with r b or r c .
  • r b and r c is determined according to the vehicle speed and the distance between the vehicle and the traffic signal, comprising: determining r b and r c
  • obtaining r b and r c includes: determining r b and r c according to the estimated vehicle exiting lane time when the risk scene is a road boundary scenario, wherein the estimated vehicle The exit time is negatively correlated with r b or r c .
  • Lane time is negatively correlated with r b or r c .
  • determining r b and r c according to the estimated vehicle exiting lane time including: Determining r b and r c , r SG3 is r b or r c in the scene of the road exiting the road, where the TLC threshold indicates the critical value of the time required for the vehicle to rush out of the road, e is a natural constant, t lc is the estimated vehicle rushing out of the lane time, y is the lateral offset of the vehicle at the road boundary, W road is the width of the road, v long is the longitudinal speed of the vehicle, and v lat is the lateral direction of the vehicle The speed, ⁇ - ⁇ d represents the angle between the heading of the vehicle and the tangential direction of the road, and r 4 is the normalized parameter.
  • obtaining r b and r c includes: determining r b and r c according to at least one of the following parameters when the risk scene is an unconventional driving scene: longitudinal acceleration of the vehicle, lateral acceleration of the vehicle And the speed of the vehicle, wherein the longitudinal acceleration of the vehicle is positively correlated with r b or r c , the lateral acceleration of the vehicle is positively correlated with r b or r c , and the speed of the vehicle is positively correlated with r b or r c .
  • An unconventional form of scene is, for example, a driving scene in which the vehicle has lateral acceleration.
  • the longitudinal acceleration of the vehicle, a lat represents the lateral acceleration of the vehicle, a threshold1 represents the threshold of the lateral acceleration in the safe driving state, a threshold2 represents the threshold of the longitudinal acceleration in the safe driving state, and v represents the speed of the vehicle, v safe speed threshold value indicates the safe driving condition, r 5 is a normalization parameter.
  • obtaining r b and r c includes: determining a driving mode and a risk scene of the vehicle in a time period, where the starting point of the time period is a first moment, and the ending point of the time period is a second moment
  • the driving mode includes an automatic driving mode and/or a manual driving mode; determining, according to a risk identification manner corresponding to the risk scene, a risk value of the driving mode in the time period, wherein the current time is after the first time, and The current moment is before the second moment.
  • the above time period may be a unit of measurement, and the risk values of the driving mode during the time period are, for example, r b and r c .
  • the vehicle selects a time period among a plurality of time periods required to travel a certain distance, and determines a corresponding time in the time period by using the known driving information and the predicted driving information at a time (ie, the current time) within the time period.
  • the risk value of the driving mode for example, the determined time period is 1 second, and the current time is 0.4 seconds.
  • the vehicle can determine the driving mode in the time period is very
  • the driving risk value is 1, that is, the probability of overspeed is 100%; if there is no overspeed in the past 0.4 seconds, and the vehicle determines that the speed will remain in the legal 0.6 seconds in the future based on the current vehicle speed and acceleration Within the driving speed, the vehicle can determine that the unconventional driving risk value in the time period is a value less than 1, that is, the probability of occurrence of overspeed is less than 100%.
  • the time period required for traveling for a certain distance is divided into a plurality of time periods, and the risk value of the driving mode in each time period is determined according to the traffic context in each time period, so that the risk can be more objectively identified.
  • the method before determining the risk value of the driving mode in the time period according to the risk identification manner corresponding to the risk scenario, the method further includes: determining a position and a motion parameter of the vehicle at the first moment, Wherein, the vehicle travels based on the manual driving mode at the first moment; predicts the position of the vehicle at the third moment according to the position and the motion parameter of the vehicle at the first moment, the third moment is later than the first moment; according to the position and the first moment The position of the three moments predicts the shadow driving mode of the vehicle from the position of the first moment to the position of the third moment, and the shadow driving mode is an automatic driving mode of the vehicle based on the position and the motion parameter predicted by the vehicle in the manual driving mode; Determining the risk value of the driving mode in the time period according to the risk identification manner corresponding to the risk scenario, including: determining r a , r b , and r c according to the risk identification manner corresponding to the risk scenario at the current time, where r a is artificial risk
  • the vehicle may predict the position of the vehicle after a period of time according to the position and the motion parameter of the vehicle at the first moment, that is, predict the position of the vehicle at the third moment, wherein the vehicle is in the manual driving state at the first moment, and then, the vehicle prediction The shadow driving mode of the vehicle from the position at the first moment to the position at the third moment.
  • the above embodiment continuously corrects the route of the automatic driving according to the position and the motion parameter in the manual driving mode, so that the route of the automatic driving mode is closer to the route of the manual driving mode, so that more objective automatic driving risk data can be acquired.
  • the present application also provides a method for determining driving risk of a vehicle, the method being applied to a vehicle having an automatic driving mode and a manual driving mode, the method comprising: determining a driving mode and a risk scene of the vehicle within a time period
  • the starting point of the time period is a first time
  • the end point of the time period is a second time
  • the driving mode includes an automatic driving mode and/or a manual driving mode
  • the current time is determined according to a risk identification manner corresponding to the risk scene.
  • the risk value of the driving mode in the time period wherein the current time is after the first time, and the current time is before the second time.
  • the vehicle selects a time period among a plurality of time periods required to travel a certain distance, and determines a corresponding time in the time period by using the known driving information and the predicted driving information at a time (ie, the current time) within the time period.
  • the risk value of the driving mode for example, the determined time period is 1 second, and the current time is 0.4 seconds.
  • the vehicle can determine the driving mode in the time period is very
  • the driving risk value is 1, that is, the probability of overspeed is 100%; if there is no overspeed in the past 0.4 seconds, and the vehicle determines that the speed will remain in the legal 0.6 seconds in the future based on the current vehicle speed and acceleration Within the driving speed, the vehicle can determine that the unconventional driving risk value in the time period is a value less than 1, that is, the probability of occurrence of overspeed is less than 100%.
  • the time period required for traveling for a certain distance is divided into a plurality of time periods, and the risk value of the driving mode in each time period is determined according to the traffic context in each time period, so that the risk can be more objectively identified.
  • the method before determining the risk value of the driving mode in the time period according to the risk identification manner corresponding to the risk scenario, the method further includes: determining a position and a motion parameter of the vehicle at the first moment, Wherein, the vehicle travels based on the manual driving mode at the first moment; predicts the position of the vehicle at the third moment according to the position and the motion parameter of the vehicle at the first moment, the third moment is later than the first moment; according to the position and the first moment The position of the three moments predicts the shadow driving mode of the vehicle from the position of the first moment to the position of the third moment, and the shadow driving mode is an automatic driving mode of the vehicle based on the position and the motion parameter predicted by the vehicle in the manual driving mode; Determining the risk value of the driving mode in the time period according to the risk identification manner corresponding to the risk scenario, including: determining r a , r b , and r c according to the risk identification manner corresponding to the risk scenario at the current time, where r a is artificial risk
  • the vehicle may predict the position of the vehicle after a period of time according to the position and the motion parameter of the vehicle at the first moment, that is, predict the position of the vehicle at the third moment, wherein the vehicle is in the manual driving state at the first moment, and then, the vehicle prediction The shadow driving mode of the vehicle from the position at the first moment to the position at the third moment.
  • the above embodiment continuously corrects the route of the automatic driving according to the position and the motion parameter in the manual driving mode, so that the route of the automatic driving mode is closer to the route of the manual driving mode, so that more objective automatic driving risk data can be acquired.
  • determining r a , r b , and r c according to the risk identification manner corresponding to the risk scenario at the current moment includes: determining, when the risk scenario is a collision scenario, r a according to at least one of the following parameters, r b and r c : the speed ⁇ v of the vehicle with respect to the collision object, the time TTC required for the collision of the vehicle with the collision object based on the motion trajectory of the collision object, the vehicle estimated based on the distance between the vehicle and the collision object brake THW is the time, and the distance between the vehicle and the colliding object, wherein, Delta] v and r a, r b r c or positive correlation, with the TTC r a, r b r c or negative correlation, the THW is r a, r b, or r c is negatively correlated, and the distance between the vehicle and the collision object is negatively correlated with r a , r b or
  • TTC is negatively correlated with r a , r b or r c ; the longer the brake is available, the smaller the probability of collision, so THW and r a , r b or r c is negatively correlated; the farther the vehicle is from the collision object, the greater the possibility that the vehicle adjusts the driving route to avoid collision, and the smaller the probability of collision occurrence, therefore, the distance between the vehicle and the collision object is r a , r b or r c negative correlation.
  • r a , r b and r c are determined according to at least one of the following parameters: a speed ⁇ v of the vehicle relative to the collision object, a vehicle estimated based on the motion trajectory of the collision object, and the collision object occurrence
  • determining r a , r b , and r c according to the risk identification manner corresponding to the risk scene at the current moment including: when the risk scene is a traffic signal scene, according to the speed of the vehicle and the vehicle to the traffic signal determining the distance r a, r b and r C, wherein r and the vehicle speed a, r b C r or positive correlation, vehicle traffic lights to a distance r a, r b C r or negative.
  • the acceleration constant, s(t) represents the distance of the vehicle to the traffic signal at time t.
  • determining r a , r b , and r c according to the risk identification manner corresponding to the risk scenario at the current moment including: when the risk scenario is a road boundary scenario, the predicted vehicle exits the lane The time determines r a , r b and r c , wherein the predicted vehicle exit lane time is inversely related to r a , r b or r c .
  • Lane time is negatively correlated with r b or r c .
  • determining, according to the estimated vehicle exiting lane time, r a , r b , and r c including: according to Determining r a , r b and r c , r SG3 is r a , r b or r c in the scene of the road exit , wherein the TLC threshold indicates a critical value of the time required for the vehicle to rush out of the road, e is a natural constant, t lc is the estimated vehicle exit time, y is the lateral offset of the vehicle at the road boundary, W road is the width of the road, and v long is the longitudinal direction of the vehicle The speed, v lat is the lateral speed of the vehicle, ⁇ - ⁇ d represents the angle between the heading of the vehicle and the tangential direction of the road, and r 4 is a normalized parameter.
  • determining r a , r b , and r c according to the risk identification manner corresponding to the risk scenario at the current moment includes: determining, when the risk scenario is an unconventional driving scenario, determining the r according to at least one of the following parameters: a , r b and r c : the longitudinal acceleration of the vehicle, the lateral acceleration of the vehicle and the speed of the vehicle, wherein the longitudinal acceleration of the vehicle is positively correlated with r a , r b or r c , the lateral acceleration of the vehicle and r a , r b Or r c is positively correlated and the speed of the vehicle is positively correlated with r a , r b or r c .
  • the longitudinal acceleration of the vehicle is positively correlated with r a , r b or r c ; the greater the lateral acceleration, the greater the probability that the vehicle will make a sharp turn, therefore, the vehicle
  • the lateral acceleration is positively correlated with r a , r b or r c ; the greater the speed of the vehicle, the greater the probability of unconventional driving such as overspeed, and therefore the speed of the vehicle is positively correlated with r a , r b or r c .
  • the method further includes: transmitting r a , r b , and r c , r a , r b , and r c to the security center for evaluating the security of the artificial driving mode and the shadow driving mode.
  • the vehicle may send the above risk data or the original data corresponding to the risk data (for example, the measured speed value of the vehicle) to the safety center for the safety center to evaluate the manual driving mode and the shadow.
  • the safety of the driving mode reduces the workload of the onboard processor.
  • the method further includes: receiving security scenario information sent by the security center, the security scenario information is used to indicate a safe driving traffic scene set of the automatic driving mode; and determining the automatic driving mode according to the security scenario information.
  • a safe driving traffic scene set if the current traffic scene does not belong to a safe driving traffic scene set, the vehicle plays and/or displays a prompt message prompting the driver to perform manual driving, and/or the vehicle performs a safe docking process.
  • the security center may identify the risk scenario according to the risk data reported by the vehicle, and update the risk identification mechanism and the risk control strategy of the vehicle by transmitting the security scenario information to the vehicle.
  • the security scenario information may be Instructing the vehicle to exclude the risk scene from the safe driving traffic scene set, and updating the safe driving traffic scene set after receiving the safety scene information, and if the current traffic scene does not belong to the updated safe driving traffic scene set, the vehicle adopts safety measures. For example, playing and/or displaying a prompt message prompting the driver to perform manual driving, and/or performing a safe docking process.
  • an automatic driving safety evaluation device which can implement the functions described in the method related to the first aspect, and the functions can be implemented by hardware or by executing corresponding software by hardware.
  • the hardware or software includes units or modules corresponding to one or more of the functions described above.
  • the above apparatus includes a transceiver and a processor configured to support the apparatus to perform the functions recited in the method of the first aspect described above.
  • the transceiver is used to support communication between the device and other devices.
  • the apparatus can also include a memory for coupling with the processor that retains the program instructions and data necessary for the apparatus.
  • a computer readable storage medium the computer program code storing computer program code, when executed by a processing unit or a processor, causing an automatic driving safety assessment device to perform the first aspect Said method.
  • a chip in which instructions are stored that, when run on an automated driving safety assessment device, cause the chip to perform the method of the first aspect.
  • a computer program product comprising: computer program code, when the computer program code is run by a communication unit or transceiver of an automated driving safety assessment device, and a processing unit or processor
  • the automatic driving safety evaluation device performs the method of the above first aspect.
  • a device for determining a driving risk of a vehicle may implement the functions involved in the method of the second aspect, and the foregoing functions may be implemented by hardware or by executing corresponding software by hardware.
  • the hardware or software includes units or modules corresponding to one or more of the functions described above.
  • the above apparatus includes a transceiver and a processor configured to support the apparatus to perform the functions recited in the method of the first aspect described above.
  • the transceiver is used to support communication between the device and other devices.
  • the apparatus can also include a memory for coupling with the processor that retains the program instructions and data necessary for the apparatus.
  • a computer readable storage medium storing computer program code for executing a device for determining a driving risk of a vehicle when the computer program code is executed by a processing unit or a processor The method described in the two aspects.
  • a chip in which instructions are stored that, when run on a device that determines the driving risk of the vehicle, cause the chip to perform the method of the second aspect.
  • a computer program product comprising: computer program code, when the computer program code is determined to be a communication unit or transceiver of the device driving the vehicle, and the processing unit or processor is running
  • the apparatus for determining the driving risk of the vehicle performs the method of the second aspect described above.
  • an automatic driving safety assessment system comprising a risk control module, a risk policy module and a safety assessment module, wherein the risk control module is configured to perform the method in any of the optional embodiments of the first aspect To determine the risk of the automatic driving mode; the risk strategy module is configured to determine a risk control strategy according to the risk of the automatic driving mode determined by the risk control module, and the risk control strategy is used to reduce or eliminate the risk of the automatic driving mode; the safety evaluation strategy And determining a security of the automatic driving mode according to a risk of the automatic driving mode determined by the risk control module.
  • an automatic driving safety evaluation system comprising a risk control module, a risk policy module and a safety evaluation module, wherein the risk control module is configured to perform the method in any of the alternative embodiments of the second aspect To determine the risk of the automatic driving mode; the risk strategy module is configured to determine a risk control strategy according to the risk of the automatic driving mode determined by the risk control module, and the risk control strategy is used to reduce or eliminate the risk of the automatic driving mode; the safety evaluation strategy And determining a security of the automatic driving mode according to a risk of the automatic driving mode determined by the risk control module.
  • Figure 1 is a flow chart of a conventional automobile product release
  • Figure 2 is a schematic diagram of various risks of manual driving and the risk of autonomous driving
  • Figure 3 is a schematic diagram of a mass production route of an autonomous vehicle
  • FIG. 4 is a schematic diagram of an automatic driving safety assessment system provided by the present application.
  • FIG. 5 is a schematic flowchart of a method for evaluating automatic driving safety provided by the present application.
  • FIG. 6 is a schematic diagram of an automatic driving safety evaluation method provided by the present application.
  • FIG. 7 is a schematic diagram of risks in a collision scenario provided by the present application.
  • FIG. 8 is a schematic diagram of risks in a traffic signal scene provided by the present application.
  • FIG. 9 is a schematic diagram of a method for determining driving risk of a vehicle provided by the present application.
  • FIG. 10 is a schematic diagram of an automatic driving control vehicle trajectory provided by the present application.
  • FIG. 11 is a schematic flow chart of a shadow driving mode provided by the present application.
  • FIG. 12 is a schematic flowchart of a method for identifying a driver's driving intention provided by the present application
  • FIG. 13 is a schematic diagram of predicting a shadow trajectory based on a shadow driving mode provided by the present application.
  • FIG. 14 is a schematic diagram of comparison between a shadow driving mode risk and a manual driving mode risk provided by the present application.
  • 16 is a schematic diagram of a safe driving scene fence provided by the present application.
  • 17 is a schematic view of a vehicle end device provided by the present application.
  • Figure 18 is a schematic view of another vehicle end device provided by the present application.
  • FIG. 19 is a schematic diagram of a cloud device provided by the present application.
  • 20 is a schematic diagram of another cloud device provided by the present application.
  • the traditional automotive product release process is shown in Figure 1, including: product planning, development, verification, release, exposure to defects, and termination (ie, discontinuation).
  • the safety level requirements of the autopilot system are similar to those in the traditional automotive electronics field, but due to the unknown hidden dangers introduced by autonomous driving and the cumulative road test mileage (million-kilometer class) introduced by the complexity of the driving scene, it is difficult to implement as shown in Figure 1.
  • the traditional release process is similar to those in the traditional automotive electronics field, but due to the unknown hidden dangers introduced by autonomous driving and the cumulative road test mileage (million-kilometer class) introduced by the complexity of the driving scene, it is difficult to implement as shown in Figure 1.
  • the traditional release process is shown in Figure 1, including: product planning, development, verification, release, exposure to defects, and termination (ie, discontinuation).
  • Figure 2 shows the unknown hazard introduced by autonomous driving.
  • the outermost circle represents all driving scenarios, where the insignificant scene (A zone) occupies the majority of the driving scene, and the existing knowledge can only pass through the traffic that has occurred.
  • C zone identifying the risk scenario of the artificial driving
  • E zone the risk of self-driving and protected manual driving
  • F zone new unknown hazards
  • the self-driving car mass production route includes two routes.
  • Route 1 Small-scale production and installation of auto-driving cars, safety evaluation methods to verify their safety, verify the safety of the autopilot system, when the safety performance of the autopilot system meets the requirements, determine whether it can enter the mass production phase.
  • Route 2 Mass production of vehicles with advanced driver assistance systems, safety assessment methods to assess their safety, when the advanced driver assistance system achieves a reliable level of safety, cars with advanced driver assistance systems can be upgraded to autonomous driving car.
  • Route 1 and Route 2 represent two routes for mass production of autonomous vehicles.
  • the vehicle identifies the risk caused by the defect of the automatic driving system through the risk identification method, updates the risk control strategy in time, and adopts risk control measures before the risk has been converted into an accident to reduce the probability of accident occurrence.
  • FIG. 4 shows an automatic driving safety assessment system provided by the present application.
  • the system 400 includes a vehicle end device 410 and a cloud device 420.
  • the vehicle end device 410 includes a sensor 411, a real-time environment sensing module 412, an autopilot module 413, a shadow driving module 414, a risk identification module 415, and a risk control module 416.
  • the cloud device 420 includes a risk policy module 421 and a security assessment module 422.
  • the sensor 411 is used to collect environmental information, and the sensor 411 can be, for example, a camera and a radar.
  • the real-time environment awareness module 412 is configured to process the environment information collected by the sensor 411 for use by subsequent modules. For example, the camera detects that there is an obstacle in front of the vehicle, and the radar also detects that there is an obstacle in front of the vehicle. In fact, the obstacle detected by the camera and the radar may be the same obstacle, and the real-time environment sensing module 412 needs to be based on The information gathered by the camera and radar determines that there are actually several obstacles in front.
  • the autopilot module 413 is configured to control the vehicle travel according to the data (ie, the perception result) output by the real-time environment sensing module 412, and output the autopilot control trajectory.
  • the shadow driving module 414 is configured to simulate the automatic driving according to the data output by the real-time environment sensing module 412 when the vehicle is manually driven, and generate a shadow trajectory.
  • the workflow of the shadow driving module 414 will be described in detail below.
  • the risk identification module 415 is configured to identify the risk scenario of the manual driving, the automatic driving, and the shadow driving, and generate the risk data.
  • the risk identification module 415 identifies the risk of the manual driving according to the sensing result output by the factual environment sensing module 412, according to the automatic driving module 413.
  • the output of the autopilot track trajectory identifies the risk of automatic driving, and the risk of shadow driving is identified based on the shadow trajectory output by the shadow driving module 414.
  • the risk policy module 421 is configured to formulate a risk control strategy for the potential risk of the identified automatic driving system, and the risk control module 416 is configured to take measures to reduce or eliminate the risk according to the risk control strategy issued by the cloud.
  • the safety assessment module 422 is for evaluating the safety of the automated driving system using the accumulated risk data.
  • FIG. 4 is a schematic description of the system 400 from the perspective of function division.
  • the functions of each module may be implemented in software form, or may be implemented in hardware form, or may be implemented in a combination of software and hardware.
  • the function of the various modules of the system 400 can be determined according to the actual situation, which is not limited in this application.
  • FIG. 4 divides the system 400 into a vehicle end device and a cloud device, but the architecture of the system 400 is not limited thereto.
  • the risk identification module 415 can be deployed in the cloud device 420, and the vehicle end device 410 does not need to perform risk identification. Processing, this can reduce the burden on the processor of the vehicle end device 410; for example, the vehicle end device 410 can deploy the risk policy module and the security evaluation module, so that even if the vehicle end device 410 is in an offline state, the vehicle end device 410 can be based on Identify the risks associated with automated driving systems to develop risk strategies and assess the safety of automated driving systems.
  • FIG. 5 illustrates a schematic flow diagram of a method for assessing the safety of automated driving.
  • S501 can be performed by sensor 411, which collects real-time information, reconstructs static and dynamic environmental data, and trajectory data of the vehicle (also referred to as "self-driving") in which sensor 411 is located.
  • S502 may be executed by the risk identification module 415 to identify risks in different vehicle control modes, wherein if the vehicle is in the automatic driving mode, the risk identification module 415 calculates the risk characteristics of the automatic driving according to the automatic driving control track; if the vehicle is in manual In the driving mode, the risk identification module 415 calculates the risk characteristics of the artificial driving according to the driver's controlled vehicle trajectory, and calculates the risk characteristics of the shadow driving according to the shadow trajectory.
  • the risk control module 416 identifies the potential defects of the automated driving system based on the risk characteristics of the automatic driving mode calculated by the risk identification module 415 and the risk characteristics of the shadow driving mode, and updates the risk control strategy.
  • S504 may be performed by a security assessment module 422 that evaluates the security of the automated driving mode based on the accumulated risk characteristics of the various driving modes, the results of which are used as a basis for security verification and mass production release.
  • FIG. 6 shows an automatic driving safety assessment method provided by the present application.
  • the method 600 includes:
  • r b is a risk value of a shadow driving mode of the vehicle in the first measurement unit
  • r c is a risk value of the automatic driving mode of the vehicle based on the preset route in the first measurement unit
  • the driving mode is a real-time route-based automatic driving mode of the vehicle
  • the real-time route is a route predicted according to the position and the motion parameter of the vehicle in the manual driving mode
  • the first unit of measurement is a time period or a distance.
  • R C is a risk value of the automatic driving mode based on the preset route of the vehicle in multiple measuring units; wherein, R B and R C are used to determine the safety of the automatic driving mode of the vehicle does it reach the requirement.
  • Method 600 can be performed by cloud device 420 or by a vehicle-end device configured with a security assessment module.
  • r b and r c reflect the probability that the risk is converted into an accident, which may be the probability value of the risk being converted into an accident, or other types of values that can quantify the risk of automatic driving.
  • the unit of measurement of r b and r c may be a time period or a distance.
  • Risk behavior can be defined as explicit or implicit driving behaviors that violate safety objectives (including manual driving, automatic driving, and shadow driving), where explicit expression of driving behavior has turned into an accident, implicitly indicating that driving behavior may be transformed For the accident. Therefore, the identification of risk behaviors includes:
  • the identification of hidden risks does not reserve sufficient safety distance with the preceding vehicle, the maximum speed of the safe driving speed is still accelerated, and the intersection does not decelerate in time; the hidden risk also includes abnormal driving behavior, these driving The behavior may cause his car to fail to react in time to cause a passive accident, such as abnormal acceleration and deceleration, and not traveling in the direction of the road.
  • the above safety goal includes but is not limited to the following definitions:
  • SG4 Avoid unconventional driving (including overspeed, abnormal acceleration and deceleration and steering)
  • the risk value when there is a dominant risk, can be defined as 1; when there is no risk, the risk value can be defined as 0; and the risk value of the recessive risk is greater than 0 and less than 1.
  • the method of calculating the risk value in each scenario is described in detail below.
  • the first unit of measurement is any unit of measurement.
  • the automatic driving mode based on the preset route is the automatic driving mode shown in FIG. 5, which controls the driving mode of the vehicle for the automatic driving module 413, and the automatic driving module 413 acquires the starting point and the ending point (ie, the preset route) of a certain distance. Control the vehicle to travel the section, the driver does not participate in driving.
  • the driving mode based on the real-time route is an automatic driving mode simulated by the shadow driving module 414 during the driver's control of the running of the vehicle, and the shadow driving module 414 continuously corrects the shadow trajectory according to the position and the motion parameter of the vehicle when the driver controls the vehicle, thereby avoiding the driver's control of the vehicle.
  • the route passes through the A street, and the automatic driving route passes through the B street, making the safety assessment of the automatic driving more objective.
  • R B may be, for example, an average value obtained by dividing the sum of r b of a plurality of measurement units by a plurality of measurement units, or may be a sum after each measurement unit of r b multiplied by a weighting coefficient.
  • the average value obtained by dividing by a plurality of units of measurement; for the same reason, R C may be, for example, an average value obtained by dividing the sum of r c of a plurality of units of measurement by a plurality of units of measurement, or may be an amount of each unit of measure. c is multiplied by the sum of the weighting coefficients and then divided by the average of the plurality of units of measure.
  • R B and R C may also be values calculated by other methods, and the present application does not limit the method of calculating R B and R C .
  • the automatic driving safety assessment method collects risk data generated by an automatic driving mode based on a preset route, in addition to collecting risk data generated by an automatic driving mode based on a preset route, and the real-time route is based on manual driving.
  • the mode of the vehicle's position and motion parameters predicted autopilot route ie, shadow trajectory
  • the shadow driving module 414 may re-plan the autopilot route based on the driver's control route and based on re-planning
  • the automatic driving route determines the safety of the automatic driving, so that the safety of the automatic driving can be more realistically evaluated.
  • the method 600 further includes:
  • the above embodiment can be implemented by means for determining R B and R C without transmitting R B and R C to other devices and evaluating the safety of the vehicle's automatic driving mode by other devices, thereby improving the real-time safety assessment of the automatic driving mode. Sex.
  • the automatic driving risk threshold indicates a threshold value of the risk value of the automatic driving mode in each unit of measurement.
  • the automatic driving risk threshold may be a value set according to the statistical result, for example, the maximum value of the average driving risk per kilometer in the mileage of one million kilometers traveled by the vehicle, if the vehicle travels one million kilometers after the R B +R If the value of C is less than or equal to the maximum value, the auto-driving function of the vehicle meets the safety requirements and can be mass-produced. If the value of R B +R C is greater than the maximum value after the vehicle travels one million kilometers, the vehicle automatically The driving function does not meet the safety requirements and cannot be mass-produced and needs to be improved.
  • R B R C + R B and R C are respectively corresponding to the mileage, for example, it may be set per million kilometers R B + R C R C corresponding to the mileage in less than 10 Ten thousand kilometers, so that the safety of autonomous driving can be evaluated more objectively and realistically.
  • the automatic driving risk threshold is a(R A +R′ A ), where a is a risk tolerance coefficient, and R A is, for example, a mode of manual driving of the vehicle in multiple units of measurement.
  • the average value of the risk value, the road segment through which the vehicle passes in a plurality of measurement units is the first road segment, and R' A is, for example, an average value of the risk values of the vehicle passing through the first road segment based on the manual driving mode outside the plurality of measurement units.
  • the risk value of the automatic driving can be compared with the risk value of the manual driving, and the tolerance coefficient of the automatic driving risk value can be accepted, so that the safety of the automatic driving can be judged whether the safety is satisfied.
  • R A is the risk value obtained after the vehicle is driven for 500,000 kilometers in the manual driving mode.
  • the road section through which the 500,000 kilometers pass is the first road section, and R' A is the vehicle in the manual driving mode after passing the first road section again.
  • the obtained risk value; R B + R C is the risk value obtained after driving one million kilometers of self-driving, and a means that the risk value of the automatic driving needs to be smaller than the risk value of the manual driving and a in the case of traveling the same length.
  • the product can be used to determine that the safety of the automatic driving mode meets the requirements, and a is usually a constant greater than 0 and less than 1.
  • determining R B from r b and determining R C from r c includes:
  • the s event reflects the severity of the accident in different risk scenarios, r i reflects the probability that the risk is converted into an accident, and ⁇ reflects the cumulative risk of multiple units of measure.
  • the s event of the collision can be set to 1, and the s event of the other object can be from 0.1- according to the nature of the object. Choose from 0.9.
  • s event can be defined based on accident statistics. When any of the above behaviors occurs, the corresponding s event is equal to 0.01.
  • X may be the mileage or time traveled by the vehicle corresponding to R i , and R i reflects the probability that the risk is converted into an accident per average unit of measure (eg, per kilometer or hour).
  • the above formula can objectively reflect the risk of autonomous driving.
  • obtaining r b and r c includes:
  • determining r b and r c according to at least one of the following parameters: a speed ⁇ v of the vehicle relative to the collision object, and a collision of the vehicle and the collision object estimated based on the motion trajectory of the collision object Time TTC, the braking time THW of the vehicle based on the distance between the vehicle and the collision object, and the distance between the vehicle and the collision object, where ⁇ v is positively correlated with r b or r c , TTC and r b or r c is negatively correlated, THW is negatively correlated with r b or r c , and the distance between the vehicle and the collision object is negatively correlated with r b or r c .
  • TTC time to collision occurrence
  • THW time to collision occurrence
  • vehicle and collision object The farther the distance is, the more likely the vehicle adjusts the travel route to avoid collisions, and the less likely the collision occurs, so the distance between the vehicle and the collision object is negatively correlated with r b or r c .
  • FIG. 7 is a schematic diagram showing the risk in a collision scenario provided by the present application.
  • the trajectory of the self-vehicle in Fig. 7 is the line segment from the state (t2, s2, v2, a2) to the state (t0, s0, v0, a0), and the trajectory of the car is from the state (t1, s1, v1, a1). )
  • the two cars face each other and collide at the intersection of the two lines shown in the figure.
  • the above two trajectories are all trajectories predicted by the vehicle, and the self-vehicle recognizes the object that intersects with the self-vehicle trajectory in a certain period of time (for example, 5 seconds) according to the sensing result of the sensor, and the collision risk value in the future period is taken from the vehicle and The maximum value of the collision risk value between the above objects.
  • TTC(t) represents the TTC at time t
  • THW(t) represents the THW at time t.
  • r 0 , r 1 and r 2 are normalized parameters respectively, v is the speed of the vehicle at time t, ⁇ v(t) is ⁇ v at time t, exp is an index based on the natural constant, and s is the vehicle and the collision object. The distance between them.
  • h(t event , obj) represents the collision risk event (obj) for a period of time (t Event ) raises a potential conflict event
  • h(t event , obj) indicates that the occurrence of a collision risk event (obj) leads to an increase in the probability of collision within a period of time (t event )
  • h(t event , obj) 0 indicates that no collision risk event has occurred.
  • each normalization parameter is used to represent the weight of each parameter. For example, each normalization parameter may take a value greater than or equal to 0 and less than or equal to 1.
  • TTC less than 1s and “THW less than 1s” are collision risk events (obj), which can predict the time of collision and select a period of time near the predicted collision time (t event) Calculate the collision risk value (the probability of collision) in the period of time, and take the largest value of the multiple collision risk values (ie, f(t) max ) in the period as the collision risk value of the period.
  • obtaining r b and r c includes:
  • FIG. 8 is a schematic diagram showing the risk in a traffic light scene provided by the present application.
  • the risk of the traffic signal scene is mainly in violation of the rules indicated by the traffic lights, and the corresponding risk value can be obtained according to the speed of the vehicle.
  • the vehicle distance traffic signal 50m, the distance s of the vehicle from the intersection stop line, the speed v of the vehicle, and the acceleration a of the vehicle constitute the current vehicle state.
  • the picture on the right shows the distance-speed curve of the vehicle.
  • the traffic signal is forbidden, the vehicle needs to decelerate and reduce the speed to 0 before stopping the line. Different accelerations form different speed-distance curves.
  • the normal acceleration refers to an acceleration whose absolute value is smaller than the absolute value of the preset acceleration, for example, The preset acceleration value is 0.4g. If the vehicle adopts an acceleration of -0.3g to stop before stopping the line, the r b or r c of the vehicle is equal to 0.
  • the self-vehicle When the data-distance curve of the self-vehicle is located between the curve 1 and the curve 2 shown in Fig. 8, the self-vehicle has a hidden risk of violating the traffic signal. At this time, the self-vehicle needs to reduce the speed to 0 before stopping the line. The larger the absolute value of the acceleration, the larger the value of r b or r c .
  • h(t event ,Light) indicates the occurrence of traffic light risk event (Light).
  • the normalized parameter is used to represent the weight of the parameter. For example, each normalized parameter may take a value greater than or equal to 0 and less than or equal to 1.
  • the traffic signal at the intersection shows the forbidden signal.
  • the speed of the vehicle is 100km/h, there is a great probability of illegal traffic lights (such as red light).
  • the above-mentioned vehicle is 50m away from the intersection.
  • the traffic light at the intersection shows the forbidden signal.
  • the speed of the vehicle is 100km/h, which is the traffic light risk event (Light). It can predict the red light moment and select the time period (t event ) near the predicted red light time.
  • the red light risk value (the probability of a red light occurrence) during the period of time, taking the largest value among the multiple red light risk values (ie, f(t) max ) during that period as the red light risk value for the period of time.
  • obtaining r b and r c includes:
  • r b and r c are determined according to the estimated vehicle exit time, wherein the estimated vehicle exit time is negatively correlated with r b or r c .
  • Lane time is negatively correlated with r b or r c .
  • r SG3 is r b or r c in the scene of the road exiting the road
  • the TLC threshold indicates the critical value of the time required for the vehicle to rush out of the road
  • e is a natural constant
  • t lc is the predicted vehicle rushing out of the lane time
  • y is the lateral offset of the vehicle at the road boundary
  • W road is the width of the road
  • v long is the longitudinal speed of the vehicle
  • v lat is the lateral direction of the vehicle Speed
  • ⁇ - ⁇ d represents the angle between the heading of the vehicle and the tangential direction of the road
  • r 4 is the normalized parameter
  • the normalized parameter is used to represent the weight of the parameter.
  • obtaining r b and r c includes:
  • determining r b and r c according to at least one of the following parameters: longitudinal acceleration of the vehicle, lateral acceleration of the vehicle, and vehicle Speed, where the longitudinal acceleration of the vehicle is positively correlated with r b or r c , the lateral acceleration of the vehicle is positively correlated with r b or r c , and the speed of the vehicle is positively correlated with r b or r c .
  • the above longitudinal acceleration refers to an acceleration in which the acceleration direction is the same or opposite to the speed direction (for example, an acceleration obtained by stepping on a throttle or braking on a straight line), and the lateral acceleration refers to an angle in which the acceleration direction and the speed direction have an angle other than zero. Acceleration (for example, acceleration that causes the vehicle to drift when cornering).
  • the longitudinal acceleration of the vehicle is positively correlated with r b or r c ; the greater the lateral acceleration, the greater the probability of a sharp turn of the vehicle. Therefore, the lateral acceleration of the vehicle is r b or r c is positively correlated; the greater the speed of the vehicle, the greater the probability of unconventional driving such as overspeed, and therefore, the speed of the vehicle is positively correlated with r b or r c .
  • the threshold of acceleration may be, for example, 0.5 g
  • a threshold 2 represents a threshold of longitudinal acceleration in a safe driving state, for example, may be 0.7 g
  • v represents the speed of the vehicle
  • v safe represents a speed threshold in a safe driving state, and may refer to the road.
  • the speed limit is determined and r 5 is the normalized parameter.
  • the automatic driving safety evaluation method provided by the present application is described in detail above. If the information received by the cloud device 420 from the vehicle end device 410 is r b or r c , the cloud device 420 does not need to perform S611 to S614 again, if the cloud device 420 The information received by the vehicle end device 410 is sensor-aware data, and the cloud device 420 also needs to execute S611 to S614 to acquire r b and r c .
  • FIG. 9 illustrates a method for determining driving risk of a vehicle provided by the present application.
  • the method 900 is applied to a vehicle having an automatic driving mode and a manual driving mode, and the method 900 includes:
  • S910 Determine a driving mode and a risk scene of the vehicle in a time period, where the starting point of the time period is a first time, and the end of the time period is a second time, the driving mode includes an automatic driving mode and/or a manual driving mode. .
  • S920 Determine, according to the risk identification manner corresponding to the risk scenario, the risk value of the driving mode in the time period, where the current time is after the first time, and the current time is before the second time.
  • the method 900 can be performed, for example, by the vehicle end device 410, which selects a time period among a plurality of time periods required for the vehicle to travel a certain distance, and utilizes a time (ie, the current time) within the time period.
  • the known driving information and the predicted driving information determine the risk value of the corresponding driving mode in the time period, for example, the determined time period is 1 second, the current time is 0.4 seconds, and if the speeding has occurred within 0.4 seconds has elapsed
  • the vehicle can determine that the unconventional driving risk value of the driving mode in the time period is 1, that is, the probability of occurrence of overspeed is 100%; if the overspeeding situation does not occur within 0.4 seconds, and the vehicle is based on the current
  • the speed and acceleration of the vehicle determine that the vehicle speed will remain within the legal travel speed in the next 0.6 seconds, and the vehicle can determine that the unconventional driving risk value for the time period is a value less than one, that is, the probability of overspeed occurring is less than 100%.
  • the embodiment divides the time period required to travel a certain distance into a plurality of time segments, and determines according to the traffic context in each time period (ie, information related to the traffic environment before the current time and after the current time) The risk value of the driving mode in each time period, so that the risk can be more objectively identified.
  • the manner of determining a time period is merely an example.
  • the vehicle end device 410 can also determine a time period by other methods, for example, determining the time required to travel a distance according to the distance and speed, which is described in S910. "One time period"; for example, the first time and the second time are determined at the current time, thereby determining "one time period" described in S910.
  • the above embodiment may determine the risk value of the automatic driving mode and/or the manual driving mode, wherein the automatic driving mode includes a shadow driving mode, and the safety of the automatic driving system based on the risk value of the shadow driving mode is the automatic driving safety provided by the present application.
  • the evaluation method can achieve an important factor in better technical effects than the prior art.
  • the purpose of the shadow driving mode is to generate the simulated autopilot track by software and hardware running in the background in the manual driving mode to identify the potential risks of the autopilot system, thereby further identifying the hidden dangers of the autopilot system and accumulating safety assessments. Required data.
  • the automatic driving control trajectory generated directly based on the sensor's sensory data cannot represent the true risk level of the automatic driving system, and may cause distortion of the automatic driving safety assessment result.
  • Figure 10 shows a schematic diagram of a self-driving control vehicle trajectory.
  • the dotted arrow indicates the actual trajectory of the vehicle in the manual driving mode
  • the solid arrow indicates the automatic driving trajectory simulated according to the sensor's sensory data. Since the sensing data indicates that there is a car in the left lane, the automatic driving system will simulate continuing along the original lane. The trajectory, in fact, the vehicle does not travel along the original lane, but has changed lanes.
  • the above-mentioned automatic driving trajectory simulated according to the sensor's sensory data cannot reflect the real driving situation, resulting in distortion of the trajectory-based automatic driving safety assessment result.
  • the method 900 further includes:
  • S901 Determine a position and a motion parameter of the vehicle at the first moment, wherein the vehicle travels based on the manual driving mode at the first moment.
  • S903 predict a shadow driving mode of the vehicle from a position of the first moment to a position of the third moment according to the position of the first moment and the position of the third moment, where the shadow driving mode is based on the position and motion parameter prediction of the vehicle in the manual driving mode.
  • the automatic driving mode of the vehicle is based on the position and motion parameter prediction of the vehicle in the manual driving mode.
  • S920 includes:
  • S921 Determine r a , r b , and r c according to a risk identification manner corresponding to the risk scene at a current moment, where r a is a risk of the artificial driving mode in the time period, and r b is a shadow driving mode in the time period
  • the risk, r c is the risk of the automatic driving mode based on the preset route during the time period.
  • the vehicle may predict the position of the vehicle after a period of time according to the position and the motion parameter of the vehicle at the first moment, that is, predict the position of the vehicle at the third moment, wherein the vehicle is in the manual driving state at the first moment, and then, the vehicle prediction The shadow driving mode of the vehicle from the position at the first moment to the position at the third moment.
  • the above embodiment continuously corrects the route of the automatic driving according to the position and the motion parameter in the manual driving mode, so that the route of the automatic driving mode is closer to the route of the manual driving mode, so that more objective automatic driving risk data can be acquired.
  • the third time may be the current time described in S920, or may be the second time, or may be a time after the second time.
  • FIG. 11 shows a flow chart of a shadow driving mode provided by the present application.
  • the shadow driving module 414 identifies the driver's driving intention, ie, the content described in S901 to S903, based on the vehicle's motion parameters (eg, speed and acceleration) at the first moment, and then identifies the driver's driving intention.
  • a shadow trajectory is generated as input information of the shadow driving mode 414, thereby improving the degree of coincidence of the shadow trajectory with the driver's controlled vehicle trajectory.
  • the driver's driving intentions described above include, but are not limited to, at least one of driving, turning, and changing lanes along the current lane.
  • FIG. 12 shows a flow chart of a method for identifying a driver's intention provided by the present application.
  • the above embodiment uses the current information to predict the information of the future time, and uses the predicted future time information to determine the driving intention of the current driver, thereby eliminating the deviation caused by the driver's intention based on the sensor's sensing data.
  • FIG. 13 is a schematic diagram of predicting a shadow trajectory based on a shadow driving mode provided by the present application.
  • the shadow trajectory shown in Figure 13 is roughly the same as the driver's trajectory, but slightly different. According to the driver's trajectory and shadow trajectory, different risk values are obtained. The difference between the risk values can be used to evaluate the artificial driving mode and the automatic driving mode. Which is more secure.
  • the driver's driving intention is a risk intention (eg, an aggressive cut-in)
  • a risk intention e.g, an aggressive cut-in
  • a safer decision is selected (eg, , keep the lane in the lane) to generate a shadow track.
  • FIG. 11 A schematic diagram of the shadow trajectory risk value calculated by the step of calculating the shadow trajectory shown in FIG. 11 is shown in FIG.
  • the solid line indicates the risk value obtained by the driver-controlled vehicle trajectory calculated in accordance with S611 to S614, and the broken line corresponds to each time segment calculated according to S611 to S614 (for example, the time period from the first time to the third time).
  • the risk value of the shadow track is the risk value obtained by the driver-controlled vehicle trajectory calculated in accordance with S611 to S614, and the broken line corresponds to each time segment calculated according to S611 to S614 (for example, the time period from the first time to the third time).
  • Example 2 the driver's car track continues to fall into the risk zone, and the shadow track returns from the risk zone to the safe zone, indicating that the risk of driving is reduced by the system's analog car control.
  • This data can be used as an autonomous driving system with better safety than the driver. An evidence of car control.
  • Example 3 both the driver's control track and the shadow-controlled track continue to fall into the risk zone, but the risk value of the shadow track is less than the driver's control, indicating that the driving control is reduced by the analog control of the automatic driving system.
  • the driving system safety performance is superior to the driver's control of the car.
  • the data in the three examples above can be used as data for the automated driving safety assessment of method 600.
  • the shadow trajectory generation method is consistent with the automatic driving control algorithm; the risk calculation method of the shadow trajectory is consistent with the automatic driving/driver driving trajectory risk calculation method, thereby ensuring the risk of the shadow trajectory.
  • the method 900 further includes:
  • the vehicle can send the above risk data to the safety center for the safety center to evaluate the safety of the manual driving mode and the shadow driving mode, and reduce the workload of the vehicle processor.
  • the above security center is, for example, a cloud device 420.
  • the risk control module 416 reduces or eliminates the risk.
  • FIG. 15 is a schematic flowchart of a method for risk control provided by the present application.
  • the risk scene data includes, but is not limited to, at least one of the sensor data acquired by the sensor 411, the trajectory generated by the autopilot module 413, and the trajectory generated by the shadow driving module 414, and the risk identification module 415 confirms according to the risk scene data described above. Whether the current scenario is a risk scenario. If the risk identification module 415 confirms that the current scenario is a non-risk scenario and discovers that the current scenario is actually a risk scenario, the risk identification mechanism may be adjusted, for example, adjusting the risk identification threshold parameter and modifying the risk identification algorithm. Wait.
  • the risk identification module 415 confirms that the current scenario is a risk scenario, and determines whether the risk scenario is caused by a hidden danger of the automatic driving system, the risk identification module 415 reports the risk scenario to the risk policy module 421, and the risk policy module 421 sends the risk policy module 421 to the risk control module 416.
  • the policy update information updates the risk control strategy of the vehicle end device 410 to exclude the risk scene from the security scene fence.
  • the method 900 further includes:
  • S940 Receive security scene information sent by the security center, where the security scenario information is used to indicate a safe driving traffic scene set of the automatic driving mode.
  • S950 Determine a safe driving traffic scene set of the automatic driving mode according to the safety scenario information.
  • S960 If the current traffic scene does not belong to the safe driving traffic scene set, the vehicle plays and/or displays prompt information prompting the driver to perform manual driving, and/or the vehicle performs safe docking processing.
  • the security center may identify the risk scenario according to the risk data reported by the vehicle, and update the risk identification mechanism and the risk control strategy of the vehicle by transmitting the security scenario information to the vehicle.
  • the security scenario information may be Instructing the vehicle to exclude the risk scene from the safe driving traffic scene set, and updating the safe driving traffic scene set after receiving the safety scene information, and if the current traffic scene does not belong to the updated safe driving traffic scene set, the vehicle adopts safety measures. For example, playing and/or displaying a prompt message prompting the driver to perform manual driving, and/or performing a safe docking process.
  • the above-mentioned safe driving traffic scene set may also be referred to as a safe driving scene fence, which is composed of an environment, a static scene, a dynamic scene, and a self-driving behavior.
  • a safe driving scene fence which is composed of an environment, a static scene, a dynamic scene, and a self-driving behavior.
  • the definition of the scene fence is as shown in FIG. 16.
  • the contents in the parentheses are the contents of the safe driving scene fence of the automatic driving system.
  • the cloud device 420 timely transmits the security scene information according to the identified hidden danger of the automatic driving system, and updates the content of the safe driving scene fence.
  • the risk control module 416 obtains the recognition result of each element of the scene fence according to the map input and the real-time perception result, and updates the surrounding content of the safe driving scene fence according to the security scene information sent by the cloud device 420, in the automatic driving module 413.
  • the risk control module 416 determines that the current scene exceeds the fence of the safe driving scene, the driver is reminded to take over, or the vehicle is controlled to be safely parked.
  • a risk event is due to the intersection width greater than 20m, resulting in the system not being able to accurately identify a certain type of traffic light, causing potential risks in the updated safe driving scene fence
  • the width of the intersection is less than 20m.
  • a risk event is due to the fact that the autopilot system is unable to measure the speed of the low speed vehicle in time, causing a potential risk, then in the updated safe driving scene fence, the dynamic scene
  • the traffic speed is applicable to 10 to 50 kilometers per hour (kph).
  • the automatic driving system detects that the current traffic is less than 10kph, the automatic driving system prompts the driver to take over the car, or gently pulls to the side to stop.
  • driver's ring strategy can be adjusted, such as adjusting the reminder interval.
  • vehicle end device and the cloud device include corresponding hardware structures and/or software modules for performing various functions in order to implement the above functions.
  • the present application can be implemented in a combination of hardware or hardware and computer software in combination with the elements and algorithm steps of the various examples described in the embodiments disclosed herein. Whether a function is implemented in hardware or computer software to drive hardware depends on the specific application and design constraints of the solution. A person skilled in the art can use different methods to implement the described functions for each particular application, but such implementation should not be considered to be beyond the scope of the present application.
  • the present application may divide the functional units of the vehicle end device and the cloud device according to the above method example.
  • each functional unit may be divided according to each function, or two or more functions may be integrated into one processing unit.
  • the above integrated unit can be implemented in the form of hardware or in the form of a software functional unit. It should be noted that the division of the unit in the present application is schematic, and is only a logical function division, and the actual implementation may have another division manner.
  • Fig. 17 shows a possible structural diagram of the vehicle end device involved in the above embodiment.
  • the vehicle end device 1700 includes a processing unit 1702 and a communication unit 1703.
  • the processing unit 1702 is for controlling management of the actions of the vehicle end device 1700, for example, the processing unit 1702 is configured to support the vehicle end device 1700 to perform S620 and/or other processes for the techniques described herein.
  • the communication unit 1703 is configured to support communication between the vehicle end device 1700 and the cloud device.
  • the vehicle end device 1700 can also include a storage unit 1701 for storing program codes and data of the vehicle end device 1700.
  • the processing unit 1702 is configured to execute S610-S630, wherein when the processing unit 1702 executes S610, the processing unit 1702 may also be referred to as the obtaining unit 1702.
  • the storage unit 1701 needs to store the intermediate data generated by the processing unit 1702 in the process of performing the above steps, and the intermediate data is, for example, r b and r c , so that the processing unit 1702 is based on the risk value of the shadow driving mode in the plurality of units of measurement and based on The risk value of the automatic driving mode of the preset route determines whether the safety of the vehicle's automatic driving mode meets the requirements.
  • the processing unit 1702 is configured to execute S610, and control the communication unit 1703 to send r b and r c to the cloud device, so that the cloud device performs S620-S630.
  • the processing unit 1702 is configured to control the communication unit 1703 to send original data (for example, sensor data) for calculating r b and r c to the cloud device, so that the cloud device performs S610-S630. .
  • the processing unit 1702 is configured to perform S910 and S920.
  • the processing unit 1702 is configured to control the communication unit 1703 to execute S930.
  • the processing unit 1702 may be a processor or a controller, such as a central processing unit (CPU), a general purpose processor, a digital signal processor (DSP), and an application-specific integrated circuit. , ASIC), field programmable gate array (FPGA) or other programmable logic device, transistor logic device, hardware component, or any combination thereof. It is possible to implement or carry out the various illustrative logical blocks, modules and circuits described in connection with the present disclosure.
  • the processor may also be a combination of computing functions, for example, including one or more microprocessor combinations, a combination of a DSP and a microprocessor, and the like.
  • the communication unit 1703 may be a transceiver or the like.
  • the storage unit 1701 may be a memory.
  • the vehicle end device referred to in the present application may be the vehicle end device 1800 shown in FIG.
  • the vehicle end device 1800 includes a processor 1802, a transceiver 1803, and a memory 1801.
  • the transceiver 1803, the processor 1802, and the memory 1801 can communicate with each other through an internal connection path to transfer control and/or data signals.
  • FIG. 19 shows a possible structural diagram of the cloud device involved in the above embodiment.
  • the cloud device 1900 includes a processing unit 1902 and a communication unit 1903.
  • the processing unit 1902 is configured to control and manage the actions of the cloud device 1900.
  • the processing unit 1902 is configured to support the cloud device 1900 to execute S620 and/or other processes for the techniques described herein.
  • the communication unit 1903 is configured to support communication between the cloud device 1900 and the vehicle end device.
  • the cloud device 1900 may further include a storage unit 1901 for storing program codes and data of the cloud device 1900.
  • the processing unit 1902 is configured to control the communication unit 1903 to execute S610, that is, receive r b and r c from the vehicle end device.
  • the communication unit 1903 may also be referred to as an obtaining unit 1903, and the processing unit 1902 is also used to execute S620 to S641.
  • the storage unit 1901 needs to store intermediate data generated by the processing unit 1902 in the process of performing the above steps, the intermediate data being, for example, r b and r c , so that the processing unit 1902 is based on the risk value of the shadow driving mode in a plurality of units of measurement and based on The risk value of the automatic driving mode of the preset route determines whether the safety of the vehicle's automatic driving mode meets the requirements.
  • the processing unit 1902 is configured to control the communication unit 1903 to receive raw data (eg, sensor data) for calculating r b and r c from the vehicle end device, and the processing unit 1902 is further configured to perform communication according to the communication.
  • the raw data received by the unit 1903 is obtained by r b and r c .
  • the processing unit 1902 may also be referred to as an obtaining unit 1902.
  • the processing unit 1902 is further configured to execute S620-S641, and the storage unit 1901 needs to store the processing unit 1902 to perform the above steps.
  • the intermediate data generated in the process is, for example, r b and r c , so that the processing unit 1902 determines the vehicle based on the risk value of the shadow driving mode in the plurality of measurement units and the risk value of the automatic driving mode based on the preset route. Whether the safety of the automatic driving mode meets the requirements.
  • the processing unit 1902 is configured to execute S910 and S920.
  • the processing unit 1902 is configured to control the communication unit 1903 to execute S930.
  • Processing unit 1902 can be a processor or controller, such as a CPU, general purpose processor, DSP, ASIC, FPGA or other programmable logic device, transistor logic device, hardware component, or any combination thereof. It is possible to implement or carry out the various illustrative logical blocks, modules and circuits described in connection with the present disclosure.
  • the processor may also be a combination of computing functions, for example, including one or more microprocessor combinations, a combination of a DSP and a microprocessor, and the like.
  • the communication unit 1903 may be a transceiver or the like.
  • the storage unit 1901 may be a memory.
  • the cloud device involved in the present application may be the cloud device 2000 shown in FIG.
  • the cloud device 2000 includes a processor 2002, a transceiver 2003, and a memory 2001.
  • the transceiver 2003, the processor 2002, and the memory 2001 can communicate with each other through an internal connection path to transfer control and/or data signals.
  • the present application also provides a system for assessing the safety of automated driving, comprising one or more of the aforementioned vehicle end devices, and one or more cloud devices.
  • the processor in the present application may be an integrated circuit chip having signal processing capabilities.
  • each step of the foregoing method embodiment may be completed by an integrated logic circuit of hardware in a processor or an instruction in a form of software.
  • the above processor may be a general purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or the like. Programming logic devices, discrete gates or transistor logic devices, discrete hardware components.
  • the methods, steps, and logical block diagrams disclosed in the embodiments of the present application can be implemented or executed.
  • the general purpose processor may be a microprocessor or the processor or any conventional processor or the like.
  • the steps of the method disclosed in the embodiments of the present application may be directly implemented by the hardware decoding processor, or may be performed by a combination of hardware and software modules in the decoding processor.
  • the software module can be located in a conventional storage medium such as random access memory, flash memory, read only memory, programmable read only memory or electrically erasable programmable memory, registers, and the like.
  • the storage medium is located in the memory, and the processor reads the information in the memory and combines the hardware to complete the steps of the above method.
  • the memory in the embodiments of the present application may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memory.
  • the non-volatile memory may be a read-only memory (ROM), a programmable read only memory (ROMM), an erasable programmable read only memory (erasable PROM, EPROM), or an electrical Erase programmable EPROM (EEPROM) or flash memory.
  • the volatile memory can be a random access memory (RAM) that acts as an external cache.
  • RAM random access memory
  • RAM random access memory
  • many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (Synchronous DRAM).
  • SDRAM double data rate synchronous DRAM
  • DDR SDRAM double data rate synchronous DRAM
  • ESDRAM enhanced synchronous dynamic random access memory
  • SLDRAM synchronously connected dynamic random access memory
  • DR RAM direct memory bus random access memory
  • the present application also provides a computer readable medium having stored thereon a computer program that, when executed by a computer, implements the functions of any of the method embodiments described above.
  • the application also provides a computer program product that, when executed by a computer, implements the functions of any of the method embodiments described above.
  • the computer program product includes one or more computer instructions.
  • the computer can be a general purpose computer, a special purpose computer, a computer network, or other programmable device.
  • the computer instructions can be stored in a computer readable storage medium or transferred from one computer readable storage medium to another computer readable storage medium, for example, the computer instructions can be from a website site, computer, server or data center Transmission to another website site, computer, server or data center via wired (eg coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (eg infrared, wireless, microwave, etc.).
  • the computer readable storage medium can be any available media that can be accessed by a computer or a data storage device such as a server, data center, or the like that includes one or more available media.
  • the usable medium may be a magnetic medium (eg, a floppy disk, a hard disk, a magnetic tape), an optical medium (eg, a high-density digital video disc (DVD)), or a semiconductor medium (eg, a solid state disk, SSD)) and so on.
  • a magnetic medium eg, a floppy disk, a hard disk, a magnetic tape
  • an optical medium eg, a high-density digital video disc (DVD)
  • DVD high-density digital video disc
  • SSD solid state disk
  • the present application also provides a processing apparatus including a processor and an interface, and the processor is configured to perform the steps described in any of the foregoing method embodiments.
  • the foregoing processing device may be a chip, and the processor may be implemented by hardware or by software.
  • the processor may be a logic circuit, an integrated circuit, etc.;
  • the processor may be a general purpose processor implemented by reading software code stored in the memory.
  • the memory may be integrated in the processor and may exist independently of the processor.
  • system and “network” are used interchangeably herein.
  • the term “and/or” in this context is merely an association describing the associated object, indicating that there may be three relationships, for example, A and / or B, which may indicate that A exists separately, and both A and B exist, respectively. B these three situations.
  • the character "/" in this article generally indicates that the contextual object is an "or" relationship.
  • B corresponding to A means that B is associated with A, and B can be determined according to A.
  • determining B from A does not mean that B is only determined based on A, and that B can also be determined based on A and/or other information.
  • the disclosed systems, devices, and methods may be implemented in other manners.
  • the device embodiments described above are merely illustrative.
  • the division of cells is only a logical function division.
  • multiple units or components may be combined or integrated. Go to another system, or some features can be ignored or not executed.
  • the mutual coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interface, device or unit, or an electrical, mechanical or other form of connection.
  • the units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, that is, may be located in one place, or may be distributed to multiple network units. Some or all of the units may be selected according to actual needs to achieve the objectives of the embodiments of the present application.
  • each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.
  • the above integrated unit can be implemented in the form of hardware or in the form of a software functional unit.
  • Computer readable media includes both computer storage media and communication media including any medium that facilitates transfer of a computer program from one location to another.
  • a storage medium may be any available media that can be accessed by a computer.
  • computer readable media may comprise RAM, ROM, EEPROM, CD-ROM or other optical disk storage, disk storage media or other magnetic storage device, or can be used for carrying or storing in the form of an instruction or data structure.
  • Any connection may suitably be a computer readable medium.
  • the software is transmitted from a website, server, or other remote source using coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwave
  • the coaxial cable , fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, wireless, and microwave are included in the fixing of the associated media.
  • a disk and a disc include a compact disc (CD), a laser disc, a compact disc, a digital versatile disc (DVD), a floppy disc, and a Blu-ray disc, wherein the disc is usually magnetically copied, and the disc is The laser is used to optically replicate the data. Combinations of the above should also be included within the scope of the computer readable media.

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Automation & Control Theory (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • Transportation (AREA)
  • Mechanical Engineering (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Tourism & Hospitality (AREA)
  • Theoretical Computer Science (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Development Economics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Quality & Reliability (AREA)
  • Operations Research (AREA)
  • Human Computer Interaction (AREA)
  • Educational Administration (AREA)
  • Game Theory and Decision Science (AREA)
  • Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • General Health & Medical Sciences (AREA)
  • Computer Security & Cryptography (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • General Engineering & Computer Science (AREA)
  • Traffic Control Systems (AREA)
  • Control Of Driving Devices And Active Controlling Of Vehicle (AREA)

Abstract

一种自动驾驶安全评估方法,包括:获取r b和r c,其中,r b为第一计量单位内车辆的影子驾驶模式的风险值,r c为第一计量单位内车辆基于预设路线的自动驾驶模式的风险值,影子驾驶模式为车辆基于实时路线的自动驾驶模式,实时路线为根据处于人工驾驶模式的车辆的位置和运动参数预测的路线,第一计量单位为时间段或路程(S610);根据r b确定R B,R B为车辆在多个计量单位内的影子驾驶模式的风险值,多个计量单位包括第一计量单位(S620);根据r c确定R C,R C为车辆在多个计量单位内的基于预设路线的自动驾驶模式的风险值;其中,R B和R C用于确定车辆的自动驾驶模式的安全性是否符合要求(S630)。能够更加真实地评估自动驾驶的安全性。

Description

自动驾驶安全评估方法、装置和系统
本申请要求于2018年03月16日提交中国专利局、申请号为201810220891.X、申请名称为“自动驾驶安全评估方法、装置和系统”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
技术领域
本申请涉及自动驾驶领域,尤其涉及一种自动驾驶安全评估方法、装置和系统。
背景技术
自动驾驶汽车又称为无人驾驶汽车,其能够通过雷达等传感器获取环境信息,并根据环境信息控制车辆行驶,减少了司机控车的时间。自动驾驶汽车具有减少交通违章等优势,然而,由于司机在自动驾驶模式中控车的时间小于司机在传统的人工驾驶模式中控车的时间,自动驾驶汽车还存在很多不确定性,为了减小乘客使用自动驾驶汽车的风险,需要在销售自动驾驶汽车之前对其安全性进行严格评估。
为了评估自动驾驶汽车的安全性,需要收集大量自动驾驶汽车的行驶数据,例如,自动驾驶汽车行驶500-600万公里产生的数据,这对于汽车制造商来说是一项非常昂贵的成本。
现有的一种自动驾驶安全评估方法是将人工驾驶路线输入模拟自动驾驶系统,分别计算人工驾驶模式和自动驾驶模式中与安全相关的因素,获得人工驾驶模式与自动驾驶模式的量化值。上述方法可以使用人工驾驶模式下的数据评估自动驾驶模式的安全性,减小了自动驾驶汽车量产的成本的周期,然而,基于模拟自动驾驶系统进行安全评估无法还原真实的司机控车场景,从而导致基于模拟自动驾驶系统得到的安全评估结果与真实的自动驾驶安全性存在较大的偏差。
发明内容
本申请提供了一种自动驾驶安全评估的方法和装置,能够减小安全评估结果与真实的自动驾驶安全性的偏差。
第一方面,提供了一种自动驾驶安全评估方法,包括:获取r b和r c,其中,r b为第一计量单位内车辆的影子驾驶模式的风险值,r c为第一计量单位内车辆基于预设路线的自动驾驶模式的风险值,影子驾驶模式为车辆基于实时路线的自动驾驶模式,所述实时路线为根据处于人工驾驶模式的车辆的位置和运动参数预测的路线,第一计量单位为时间段或路程;根据r b确定R B,R B为车辆在多个计量单位内的影子驾驶模式的风险值,多个计量单位包括第一计量单位;根据r c确定R C,R C为车辆在多个计量单位内的基于预设路线的自动驾驶模式的风险值;其中,R B和R C用于确定所述车辆的自动驾驶模式的安全性是否符合要求。
r b和r c反映了风险转化为事故的概率,本申请提供的自动驾驶安全评估方法除了收集基于预设路线的自动驾驶模式产生的风险数据外,还收集基于实时路线的自动驾驶模式产生的风险数据,该实时路线为根据处于人工驾驶模式的车辆的位置和运动参数预测的路线,例如,当司机控车改变预设路线时,车载安全评估装置可以根据司机控车路线重新规划自动驾驶路线,并基于重新规划的自动驾驶路线确定自动驾驶的安全性,从而能够更加真实地评估自动驾驶的安全性。
作为一个可选的实施方式,所述方法还包括:计算R B与R C之和;当R B与R C之和小于或等于自动驾驶风险阈值时,确定车辆的自动驾驶模式的安全性符合要求。
上述实施例可以由确定R B和R C的装置实施,无需将R B和R C发送至其它设备以及通过其它设备评估车辆的自动驾驶模式的安全性,从而可以提高自动驾驶模式安全评估的实时性。
作为一个可选的实施方式,所述自动驾驶风险阈值为a(R A+R′ A),其中,a为风险容忍系数,R A为车辆在多个计量单位内的人工驾驶的模式的风险值,车辆在多个计量单位内通过的路段为第一路段,R′ A为车辆在多个计量单位之外基于人工驾驶模式通过第一路段的风险值。
可以将自动驾驶的风险值与人工驾驶的风险值做对比,并设定自动驾驶的风险值能够被接受的容忍系数,从而可以判断自动驾驶的安全性是否满足要求。
作为一个可选的实施方式,根据r b确定R B,以及根据r c确定R C,包括:根据
Figure PCTCN2019072074-appb-000001
确定R B或R C,其中,event表示风险场景,s event表示该风险场景中事故的严重程度的系数,X表示所述多个计量单位,当r i表示r b时,R i表示R B,当r i表示r c时,R i表示R C
s event反映了不同的风险场景中事故的严重程度,例如,在碰撞场景中,撞人的严重程度远高于撞其它物体的严重程度,因此,可以将撞人的严重程度设置为1,撞其它物体的严重程度例如可以根据物体的性质从0.1-0.9中选择;r i反映了风险转化为事故的概率,∑则反映了多个计量单位的风险的累计,R i则反映了平均每个计量单位内风险转化为事故的概率。上述公式可以客观地反映自动驾驶的风险。
作为一个可选的实施方式,获取r b和r c,包括:当风险场景为碰撞场景时,根据下列参数中的至少一个确定r b和r c:车辆相对于碰撞物的速度Δv、基于碰撞物的运动轨迹预估的车辆与碰撞物发生碰撞所需的时间(time to collision,TTC)、基于车辆与碰撞物之间的距离预估的车辆的刹车时间(time to headway,THW)、以及车辆与碰撞物之间的距离,其中,Δv与r b或r c正相关,TTC与r b或r c负相关,THW与r b或r c负相关,车辆与碰撞物之间的距离与r b或r c负相关。
在碰撞场景中,相对运动速度越大,风险转化为事故的概率越大,即,Δv与r b或r c正相关;距离碰撞的时间越长,车辆调整行驶路线避免碰撞的可能性越大,碰撞发生的概率越小,因此,TTC与r b或r c负相关;刹车可用时间越长,碰撞发生的概率越小,因此,THW与r b或r c负相关;车辆与碰撞物的距离越远,车辆调整行驶路线避免碰撞的可能性越大,碰撞发生的概率越小,因此,车辆与所述碰撞物之间的距离与r b或r c负相关。
作为一个可选的实施方式,根据下列参数中的至少一个确定r b和r c:车辆相对于碰撞 物的速度Δv、基于碰撞物的运动轨迹预估的车辆与碰撞物发生碰撞所需的时间TTC、基于车辆与碰撞物之间的距离预估的车辆的刹车时间THW、以及车辆与碰撞物之间的距离,包括:根据r SG1=f(t) max确定r b和r c,r SG1为碰撞场景中的r b或r c,其中,t∈{t event,h(t event,obj)=1},f(t)=Δv(t)·exp(-min{r 0TTC(t),r 1THW(t),r 2s}),{t event,h(t event,obj)=1}表示发生碰撞风险的时间段,TTC(t)表示t时刻的TTC,THW(t)表示t时刻的THW,
Figure PCTCN2019072074-appb-000002
r 0、r 1和r 2分别为归一化参数,v为t时刻车辆的速度,Δv(t)为t时刻的Δv,exp为以自然常数为底的指数,s表示车辆与碰撞物之间的距离。
作为一个可选的实施方式,获取r b和r c,包括:当风险场景为交通信号灯场景时,根据车辆的速度以及车辆到交通信号灯的距离确定r b和r c,其中,车辆的速度与r b或r c正相关,车辆到交通信号灯的距离与r b或r c负相关。
在交通信号灯场景中,车辆的速度越高,违反交通信号灯(例如闯红灯)的概率越大,因此,车辆的速度与r b或r c正相关;车辆到交通信号灯的距离越远,用于车辆改变行驶状态的时间越长,违反交通信号灯的概率越小,因此,车辆到交通信号灯的距离与r b或r c负相关。
作为一个可选的实施方式,根据车辆的速度以及车辆与交通信号灯的距离确定r b和r c,包括:根据r SG2=f(t) max确定r b和r c,r SG2为交通信号灯场景中的r b或r c,其中,t∈{t event,h(t event,Light)=1},
Figure PCTCN2019072074-appb-000003
r 3为归一化参数,{t event,h(t event,Light)=1}表示交发生违反交通信号灯风险的时间段,v(t)表示车辆在t时刻的速度,g为重力加速度常数,s(t)表示t时刻车辆到交通信号灯的距离。
作为一个可选的实施方式,获取r b和r c,包括:当风险场景为冲出道路边界场景时,根据预估的车辆冲出车道时间确定r b和r c,其中,预估的车辆冲出车道时间与r b或r c负相关。
在冲出道路边界场景中,预估的车辆冲出车道的时间越长,用于车辆调整行驶状态的时间越长,车辆冲出道路的概率也就越小,因此,预估的车辆冲出车道时间与r b或r c负相关。
作为一个可选的实施方式,根据预估的车辆冲出车道时间确定r b和r c,包括:根据
Figure PCTCN2019072074-appb-000004
确定r b和r c,r SG3为冲出道路边界场景中的r b或r c,其中,TLC threshold表示车辆冲出道路所需时间的临界值,
Figure PCTCN2019072074-appb-000005
e为自然常数,t lc为预估的车辆冲出车道时间,y为车辆在道路边界的横向偏移量,W road为道路的宽度, v long为车辆的纵向速度,v lat为车辆的横向速度,φ-φ d表示车辆的航向与道路切线方向的夹角,r 4为归一化参数。
作为一个可选的实施方式,获取r b和r c,包括:当风险场景为非常规行驶场景时,根据下列参数中的至少一个确定r b和r c:车辆的纵向加速度、车辆的横向加速度以及车辆的速度,其中,车辆的纵向加速度与r b或r c正相关,车辆的横向加速度与r b或r c正相关,车辆的速度与r b或r c正相关。
非常规形式场景例如是车辆具有横向加速度的行驶场景。纵向加速度越大,车辆超速或者急刹车的概率越大,因此,车辆的纵向加速度与r b或r c正相关;横向加速度越大,车辆急转弯的概率越大,因此,车辆的横向加速度与r b或r c正相关;车辆的速度越大,出现超速等非常规行驶的概率越大,因此,车辆的速度与r b或r c正相关。
作为一个可选的实施方式,根据下列参数中的至少一个确定r b和r c:车辆的纵向加速度、车辆的横向加速度以及车辆的速度,包括:根据r SG4=r 5确定r b和r c,r SG4为非常规行驶场景中的r b或r c,其中,(|a long|>a threshold1)||(|a lat|>a threshold2)||(v≥v safe),a long表示所述车辆的纵向加速度,a lat表示所述车辆的横向加速度,a threshold1表示安全行驶状态下横向加速度的阈值,a threshold2表示安全行驶状态下纵向加速度的阈值,v表示所述车辆的速度,v safe表示安全行驶状态下的速度阈值,r 5为归一化参数。
作为一个可选的实施方式,获取r b和r c,包括:确定一个时间段内车辆的驾驶模式和风险场景,该时间段的起始点为第一时刻,该时间段的终点为第二时刻,所述驾驶模式包括自动驾驶模式和/或人工驾驶模式;在当前时刻根据风险场景对应的风险识别方式确定所述时间段内驾驶模式的风险值,其中,当前时刻位于第一时刻之后,且当前时刻位于第二时刻之前。
上述一个时间段可以是一个计量单位,所述时间段内驾驶模式的风险值例如是r b和r c。车辆在行驶一段路程所需的多个时间段中选取一个时间段,并在该时间段内的一个时刻(即,当前时刻)利用已知的行驶信息和预测的行驶信息确定该时间段内相应的驾驶模式的风险值,例如,确定的时间段为1秒,当前时刻为0.4秒,若已过去的0.4秒内出现了超速行驶的状况,则车辆可以确定该时间段内的驾驶模式的非常规驾驶风险值为1,即,发生超速的概率为100%;若已过去的0.4秒内未出现超速行驶的状况,且车辆根据当前车辆的速度以及加速度确定未来0.6秒内车速将保持在法定行驶速度内,则车辆可以确定该时间段内的非常规驾驶风险值为一个小于1的数值,即,发生超速的概率小于100%。本实施例将行驶一段路程所需的时间段分割为多个时间段,并根据每个时间段中的交通上下文确定每个时间段内的驾驶模式的风险值,从而能够更加客观地识别风险。
作为一个可选的实施方式,在当前时刻根据风险场景对应的风险识别方式确定所述时间段内驾驶模式的风险值之前,所述方法还包括:确定车辆在第一时刻的位置和运动参数,其中,车辆在第一时刻基于人工驾驶模式行驶;根据车辆在第一时刻的位置和运动参数预测车辆在第三时刻的位置,第三时刻晚于第一时刻;根据第一时刻的位置和第三时刻的位置预测车辆从第一时刻的位置到第三时刻的位置的影子驾驶模式,影子驾驶模式为基于车辆处于人工驾驶模式下的位置和运动参数预测的车辆的自动驾驶模式;在当前时刻根据风险场景对应的风险识别方式确定所述时间段内驾驶模式的风险值,包括:在当前时刻根据 风险场景对应的风险识别方式确定r a、r b和r c,其中,r a为所述时间段内人工驾驶模式的风险值,r b为所述时间段内影子驾驶模式的风险值,r c为所述时间段内基于预设路线的自动驾驶模式的风险值。
车辆可以根据该车辆在第一时刻的位置和运动参数预测未来一段时间后车辆的位置,即,预测车辆在第三时刻的位置,其中,车辆在第一时刻处于人工驾驶状态,随后,车辆预测该车辆从第一时刻的位置到第三时刻的位置的影子驾驶模式。上述实施例根据人工驾驶模式下的位置和运动参数不断修正自动驾驶的路线,使得自动驾驶模式的路线更加接近人工驾驶模式的路线,从而可以获取更加客观地自动驾驶风险数据。
第二方面,本申请还提供了一种确定车辆的驾驶风险的方法,该方法应用于具有自动驾驶模式和人工驾驶模式的车辆,该方法包括:确定一个时间段内车辆的驾驶模式和风险场景,该时间段的起始点为第一时刻,该时间段的终点为第二时刻,所述驾驶模式包括自动驾驶模式和/或人工驾驶模式;在当前时刻根据风险场景对应的风险识别方式确定所述时间段内驾驶模式的风险值,其中,当前时刻位于第一时刻之后,且当前时刻位于第二时刻之前。
车辆在行驶一段路程所需的多个时间段中选取一个时间段,并在该时间段内的一个时刻(即,当前时刻)利用已知的行驶信息和预测的行驶信息确定该时间段内相应的驾驶模式的风险值,例如,确定的时间段为1秒,当前时刻为0.4秒,若已过去的0.4秒内出现了超速行驶的状况,则车辆可以确定该时间段内的驾驶模式的非常规驾驶风险值为1,即,发生超速的概率为100%;若已过去的0.4秒内未出现超速行驶的状况,且车辆根据当前车辆的速度以及加速度确定未来0.6秒内车速将保持在法定行驶速度内,则车辆可以确定该时间段内的非常规驾驶风险值为一个小于1的数值,即,发生超速的概率小于100%。本实施例将行驶一段路程所需的时间段分割为多个时间段,并根据每个时间段中的交通上下文确定每个时间段内的驾驶模式的风险值,从而能够更加客观地识别风险。
作为一个可选的实施方式,在当前时刻根据风险场景对应的风险识别方式确定所述时间段内驾驶模式的风险值之前,所述方法还包括:确定车辆在第一时刻的位置和运动参数,其中,车辆在第一时刻基于人工驾驶模式行驶;根据车辆在第一时刻的位置和运动参数预测车辆在第三时刻的位置,第三时刻晚于第一时刻;根据第一时刻的位置和第三时刻的位置预测车辆从第一时刻的位置到第三时刻的位置的影子驾驶模式,影子驾驶模式为基于车辆处于人工驾驶模式下的位置和运动参数预测的车辆的自动驾驶模式;在当前时刻根据风险场景对应的风险识别方式确定所述时间段内驾驶模式的风险值,包括:在当前时刻根据风险场景对应的风险识别方式确定r a、r b和r c,其中,r a为所述时间段内人工驾驶模式的风险值,r b为所述时间段内影子驾驶模式的风险值,r c为所述时间段内基于预设路线的自动驾驶模式的风险值。
车辆可以根据该车辆在第一时刻的位置和运动参数预测未来一段时间后车辆的位置,即,预测车辆在第三时刻的位置,其中,车辆在第一时刻处于人工驾驶状态,随后,车辆预测该车辆从第一时刻的位置到第三时刻的位置的影子驾驶模式。上述实施例根据人工驾驶模式下的位置和运动参数不断修正自动驾驶的路线,使得自动驾驶模式的路线更加接近人工驾驶模式的路线,从而可以获取更加客观地自动驾驶风险数据。
作为一个可选的实施方式,在当前时刻根据风险场景对应的风险识别方式确定r a、r b 和r c,包括:当风险场景为碰撞场景时,根据下列参数中的至少一个确定r a、r b和r c:车辆相对于碰撞物的速度Δv、基于碰撞物的运动轨迹预估的车辆与碰撞物发生碰撞所需的时间TTC、基于车辆与碰撞物之间的距离预估的车辆的刹车时间THW、以及车辆与碰撞物之间的距离,其中,Δv与r a、r b或r c正相关,TTC与r a、r b或r c负相关,THW与r a、r b或r c负相关,车辆与碰撞物之间的距离与r a、r b或r c负相关。
在碰撞场景中,相对运动速度越大,风险转化为事故的概率越大,即,Δv与r a、r b或r c正相关;距离碰撞的时间越长,车辆调整行驶路线避免碰撞的可能性越大,碰撞发生的概率越小,因此,TTC与r a、r b或r c负相关;刹车可用时间越长,碰撞发生的概率越小,因此,THW与r a、r b或r c负相关;车辆与碰撞物的距离越远,车辆调整行驶路线避免碰撞的可能性越大,碰撞发生的概率越小,因此,车辆与碰撞物之间的距离与r a、r b或r c负相关。
作为一个可选的实施方式,根据下列参数中的至少一个确定r a、r b和r c:车辆相对于碰撞物的速度Δv、基于碰撞物的运动轨迹预估的车辆与所述碰撞物发生碰撞所需的时间TTC、基于车辆与碰撞物之间的距离预估的车辆的刹车时间THW、以及车辆与碰撞物之间的距离,包括:根据r SG1=f(t) max确定r a、r b和r c,r SG1为碰撞场景中的r a、r b或r c,其中,t∈{t event,h(t event,obj)=1},f(t)=Δv(t)·exp(-min{r 0TTC(t),r 1THW(t),r 2s}),{t event,h(t event,obj)=1}表示发生碰撞风险的时间段,TTC(t)表示t时刻的TTC,THW(t)表示t时刻的THW,
Figure PCTCN2019072074-appb-000006
r 0、r 1和r 2分别为归一化参数,v为t时刻车辆的速度,Δv(t)为t时刻的Δv,exp为以自然常数为底的指数,s表示车辆与碰撞物之间的距离。
作为一个可选的实施方式,在当前时刻根据风险场景对应的风险识别方式确定r a、r b和r c,包括:当风险场景为交通信号灯场景时,根据车辆的速度以及车辆到交通信号灯的距离确定r a、r b和r c,其中,车辆的速度与r a、r b或r c正相关,车辆到交通信号灯的距离与r a、r b或r c负相关。
在交通信号灯场景中,车辆的速度越高,违反交通信号灯(例如闯红灯)的概率越大,因此,车辆的速度与r a、r b和r c正相关;车辆到交通信号灯的距离越远,用于车辆改变行驶状态的时间越长,违反交通信号灯的概率越小,因此,车辆到交通信号灯的距离与r a、r b和r c负相关。
作为一个可选的实施方式,所述根据所述车辆的速度以及所述车辆与交通信号灯的距离确定r a、r b和r c,包括:根据r SG2=f(t) max确定r a、r b和r c,r SG2为交通信号灯场景中的r a、r b或r c,其中,t∈{t event,h(t event,Light)=1},
Figure PCTCN2019072074-appb-000007
r 3为归一化参数,{t event,h(t event,Light)=1}表示交发生违反交通信号灯风险的时间段,v(t)表示所述车辆在t时刻的速度,g为重力加速度常数,s(t)表示t时刻所述车辆到交通信号 灯的距离。
作为一个可选的实施方式,在当前时刻根据风险场景对应的风险识别方式确定r a、r b和r c,包括:当风险场景为冲出道路边界场景时,根据预估的车辆冲出车道时间确定r a、r b和r c,其中,所述预估的所述车辆冲出车道时间与r a、r b或r c负相关。
在冲出道路边界场景中,预估的车辆冲出车道的时间越长,用于车辆调整行驶状态的时间越长,车辆冲出道路的概率也就越小,因此,预估的车辆冲出车道时间与r b或r c负相关。
作为一个可选的实施方式,所述根据预估的所述车辆冲出车道时间确定r a、r b和r c,包括:根据
Figure PCTCN2019072074-appb-000008
确定r a、r b和r c,r SG3为冲出道路边界场景中的r a、r b或r c,其中,TLC threshold表示所述车辆冲出道路所需时间的临界值,
Figure PCTCN2019072074-appb-000009
e为自然常数,t lc为所述预估的所述车辆冲出车道时间,y为所述车辆在道路边界的横向偏移量,W road为道路的宽度,v long为所述车辆的纵向速度,v lat为所述车辆的横向速度,φ-φ d表示所述车辆的航向与道路切线方向的夹角,r 4为归一化参数。
作为一个可选的实施方式,在当前时刻根据风险场景对应的风险识别方式确定r a、r b和r c,包括:当风险场景为非常规行驶场景时,根据下列参数中的至少一个确定r a、r b和r c:车辆的纵向加速度、车辆的横向加速度以及车辆的速度,其中,车辆的纵向加速度与r a、r b或r c正相关,车辆的横向加速度与r a、r b或r c正相关,车辆的速度与r a、r b或r c正相关。
纵向加速度越大,车辆超速或者急刹车的概率越大,因此,车辆的纵向加速度与r a、r b或r c正相关;横向加速度越大,车辆急转弯的概率越大,因此,车辆的横向加速度与r a、r b或r c正相关;车辆的速度越大,出现超速等非常规行驶的概率越大,因此,车辆的速度与r a、r b或r c正相关。
作为一个可选的实施方式,根据下列参数中的至少一个确定r a、r b和r c:车辆的纵向加速度、车辆的横向加速度以及车辆的速度,包括:根据r SG4=r 5确定r a、r b和r c,r SG4为非常规行驶场景中的r a、r b或r c,其中,(|a long|>a threshold1)||(|a lat|>a threshold2)||(v≥v safe),a long表示车辆的纵向加速度,a lat表示车辆的横向加速度,a threshold1表示安全行驶状态下横向加速度的阈值,a threshold2表示安全行驶状态下纵向加速度的阈值,v表示车辆的速度,v safe表示安全行驶状态下的速度阈值,r 5为归一化参数。
作为一个可选的实施方式,所述方法还包括:向安全中心发送r a、r b和r c,r a、r b和r c用于评估人工驾驶模式和影子驾驶模式的安全性。
车辆获取r a、r b或r c后可以向安全中心发送上述风险数据或者上述风险数据对应的原始数据(例如,传感器测得的车辆的速度值),用于安全中心评估人工驾驶模式和影子驾驶模式的安全性,减小了车载处理器的工作负担。
作为一个可选的实施方式,所述方法还包括:接收安全中心发送的安全场景信息,所 述安全场景信息用于指示自动驾驶模式的安全驾驶交通场景集合;根据安全场景信息确定自动驾驶模式的安全驾驶交通场景集合;若当前交通场景不属于安全驾驶交通场景集合,车辆播放和/或显示提示司机进行人工驾驶的提示信息,和/或,车辆执行安全停靠处理。
安全中心可以根据车辆上报的风险数据识别风险场景,通过向车辆发送安全场景信息更新车辆的风险识别机制和风险控制策略,例如,若风险是由自动驾驶系统的隐患导致的,则安全场景信息可以指示车辆将风险场景排除在安全驾驶交通场景集合之外,车辆接收到安全场景信息后更新安全驾驶交通场景集合,若当前交通场景不属于更新后的安全驾驶交通场景集合,则车辆采取安全措施,例如,播放和/或显示提示司机进行人工驾驶的提示信息,和/或,执行安全停靠处理。
第三方面,提供了一种自动驾驶安全评估装置,该装置可以实现上述第一方面所涉及的方法中所述的功能,上述功能可以通过硬件实现,也可以通过硬件执行相应的软件实现。所述硬件或软件包括与一个或多个上述功能相应的单元或模块。
在一种可能的设计中,上述装置包括收发器和处理器,该处理器被配置为支持该装置执行上述第一方面所涉及的方法中所述的功能。该收发器用于支持该装置与其它装置之间的通信。该装置还可以包括存储器,该存储器用于与处理器耦合,其保存该装置必要的程序指令和数据。
第四方面,提供了一种计算机可读存储介质,该计算机可读存储介质中存储了计算机程序代码,该计算机程序代码被处理单元或处理器执行时,使得自动驾驶安全评估装置执行第一方面所述的方法。
第五方面,提供了一种芯片,其中存储有指令,当其在自动驾驶安全评估装置上运行时,使得该芯片执行第一方面所述的方法。
第六方面,提供了一种计算机程序产品,该计算机程序产品包括:计算机程序代码,当该计算机程序代码被自动驾驶安全评估装置的通信单元或收发器、以及处理单元或处理器运行时,使得自动驾驶安全评估装置执行上述第一方面的方法。
第七方面,提供了一种确定车辆的驾驶风险的装置,该装置可以实现上述第二方面的方法中所涉及的功能,上述功能可以通过硬件实现,也可以通过硬件执行相应的软件实现。所述硬件或软件包括与一个或多个上述功能相应的单元或模块。
在一种可能的设计中,上述装置包括收发器和处理器,该处理器被配置为支持该装置执行上述第一方面所涉及的方法中所述的功能。该收发器用于支持该装置与其它装置之间的通信。该装置还可以包括存储器,该存储器用于与处理器耦合,其保存该装置必要的程序指令和数据。
第八方面,提供了一种计算机可读存储介质,该计算机可读存储介质中存储了计算机程序代码,该计算机程序代码被处理单元或处理器执行时,使得确定车辆的驾驶风险的装置执行第二方面所述的方法。
第九方面,提供了一种芯片,其中存储有指令,当其在确定车辆的驾驶风险的装置上运行时,使得该芯片执行第二方面所述的方法。
第十方面,提供了一种计算机程序产品,该计算机程序产品包括:计算机程序代码,当该计算机程序代码被确定车辆的驾驶风险的装置的通信单元或收发器、以及处理单元或处理器运行时,使得确定车辆的驾驶风险的装置执行上述第二方面的方法。
第十一方面,提供了一种自动驾驶安全评估系统,该系统包括风险控制模块、风险策略模块和安全评估模块,风险控制模块用于执行第一方面的任一可选的实施方式中的方法,以确定自动驾驶模式的风险;该风险策略模块用于根据风险控制模块确定的自动驾驶模式的风险确定风险控制策略,风险控制策略用于减小或消除自动驾驶模式的风险;该安全评估策略用于根据所述风险控制模块确定的所述自动驾驶模式的风险评估所述自动驾驶模式的安全性。
第十二方面,提供了一种自动驾驶安全评估系统,该系统包括风险控制模块、风险策略模块和安全评估模块,风险控制模块用于执行第二方面的任一可选的实施方式中的方法,以确定自动驾驶模式的风险;该风险策略模块用于根据风险控制模块确定的自动驾驶模式的风险确定风险控制策略,风险控制策略用于减小或消除自动驾驶模式的风险;该安全评估策略用于根据所述风险控制模块确定的所述自动驾驶模式的风险评估所述自动驾驶模式的安全性。
附图说明
图1是传统的汽车产品发布流程图;
图2是各种人工驾驶的风险和自动驾驶的风险的示意图;
图3是自动驾驶汽车量产路线示意图;
图4是本申请提供的自动驾驶安全评估系统的示意图;
图5是本申请提供的评估自动驾驶安全性的方法的示意性流程图;
图6是本申请提供的一种自动驾驶安全评估方法的示意图;
图7是本申请提供的一种碰撞场景中的风险示意图;
图8是本申请提供的一种交通信号灯场景中的风险示意图;
图9是本申请提供的一种确定车辆的驾驶风险的方法的示意图;
图10是本申请提供的一种自动驾驶控车轨迹的示意图;
图11是本申请提供的一种影子驾驶模式的示意性流程图;
图12是本申请提供的一种识别司机驾驶意图的方法的示意性流程图;
图13是本申请提供的一种基于影子驾驶模式预测影子轨迹的示意图;
图14是本申请提供的一种影子驾驶模式风险和人工驾驶模式风险的对比示意图;
图15是本申请提供的一种风险控制方法的示意性流程图;
图16是本申请提供的一种安全驾驶场景围栏的示意图;
图17是本申请提供的一种车端装置的示意图;
图18是本申请提供的另一种车端装置的示意图;
图19是本申请提供的一种云端装置的示意图;
图20是本申请提供的另一种云端装置的示意图。
具体实施方式
下面将结合附图,对本申请中的技术方案进行描述。
传统的汽车产品发布流程如图1所示,包括:产品规划、开发、验证、发布、是否暴露缺陷以及结束(即,停产)。自动驾驶系统的安全等级要求和传统汽车电子领域要求类 似,但由于自动驾驶引入的未知隐患、驾驶场景复杂性引入的累计路测里程(百万公里级)要求,导致其难以实施图1所示的传统发布流程。
图2示出了自动驾驶引入的未知隐患,最外层的圆表示所有驾驶场景,其中无关紧要的场景(A区)占据驾驶场景的绝大部分比例,现有知识仅能通过对已发生交通事故(C区)的分析,识别人工驾驶的风险场景(B区),以及自动驾驶可防护的人工驾驶的风险(E区),对自动驾驶引入的新型未知隐患(F区)没有积累足够充分的认识,因此在自动驾驶系统真正投入使用前,难以对其进行客观的安全评估。
为了促进自动驾驶汽车尽早量产以及自动驾驶汽车行业的发展,一些组织提出了自动驾驶汽车量产路线图。如图3所示,自动驾驶汽车量产路线包括两种路线。
路线1:小批量生产安装自动驾驶系统的汽车,通过安全评估方法评估其安全性,验证自动驾驶系统的安全性,当自动驾驶系统的安全性能符合要求时,确定能否进入大批量生产阶段。
路线2:大批量生产安装高级辅助驾驶系统的汽车,通过安全评估方法评估其安全性,当高级辅助驾驶系统的功能达到可靠的安全等级时,安装有高级辅助驾驶系统的汽车可以升级为自动驾驶汽车。
路线1和路线2分别代表了大批量生产自动驾驶汽车的两种路线。在自动驾驶汽车行驶的过程中,车辆通过风险识别方法识别自动驾驶系统缺陷导致的风险,及时更新风险控制策略,在风险还未转化为事故前采取风险控制措施,降低事故发生的概率。
下面,将描述本申请提供的评估自动驾驶安全性的方法、装置和系统。
图4示出了本申请提供的一种自动驾驶安全评估系统。该系统400包括车端装置410和云端装置420。
车端装置410包括传感器411、实时环境感知模块412、自动驾驶模块413、影子驾驶模块414、风险识别模块415和风险控制模块416。
云端装置420包括风险策略模块421和安全评估模块422。
传感器411用于采集环境信息,传感器411例如可以是摄像机和雷达。
实时环境感知模块412用于对传感器411采集到的环境信息进行处理,以便于后续模块使用。例如,摄像机检测到车辆正前方有一个障碍物,雷达也检测到车辆正前方有一个障碍物,实际上,摄像机和雷达检测到的障碍物可能是同一个障碍物,实时环境感知模块412需要根据摄像机和雷达采集到的信息确定前方实际有几个障碍物。
自动驾驶模块413用于根据实时环境感知模块412输出的数据(即,感知结果)控制车辆行驶,输出自动驾驶控车轨迹。
影子驾驶模块414用于在人工驾驶汽车时根据实时环境感知模块412输出的数据模拟自动驾驶,生成影子轨迹,下文将详细介绍影子驾驶模块414的工作流程。
风险识别模块415用于识别人工驾驶、自动驾驶、影子驾驶的风险场景,生成风险数据,其中,风险识别模块415根据事实环境感知模块412输出的感知结果识别人工驾驶的风险,根据自动驾驶模块413输出的自动驾驶控车轨迹识别自动驾驶的风险,根据影子驾驶模块414输出的影子轨迹识别影子驾驶的风险。
风险策略模块421用于针对已识别的自动驾驶系统的潜在风险制定风险控制策略,风险控制模块416则用于根据云端发布的风险控制策略采取措施减小或消除风险。
安全评估模块422用于利用累积的风险数据对自动驾驶系统的安全性进行评估。
图4是从功能划分的角度上对系统400进行了示例性描述,各个模块的功能在实现时可能以软件形式实现,也可能以硬件形式实现,还可能以软硬结合的形式实现,具体通过哪种形式实现系统400的各个模块的功能可根据实际情况确定,本申请对此不作限定。
不失一般性地,图4将系统400划分为车端装置和云端装置,但系统400的架构不限于此,例如,风险识别模块415可以部署在云端装置420,车端装置410无需执行风险识别处理,这样可以减轻车端装置410的处理器的负担;又例如,车端装置410可以部署风险策略模块和安全评估模块,这样,即使车端装置410处于离线状态,车端装置410也可以根据已识别的自动驾驶系统导致的风险制定风险策略,并评估自动驾驶系统的安全性。
基于系统400,图5示出了本申请提供一种评估自动驾驶安全性的方法的示意性流程图。
该方法500中,S501可以由传感器411执行,传感器411采集实时信息,重建静态和动态环境数据,以及传感器411所在车辆(也可称为“自车”)的轨迹数据。
S502可以由风险识别模块415执行,识别不同控车模式下的风险,其中,若自车处于自动驾驶模式,则风险识别模块415根据自动驾驶控车轨迹计算自动驾驶的风险特征;若车辆处于人工驾驶模式,则风险识别模块415根据司机控车轨迹计算人工驾驶的风险特征,以及根据影子轨迹计算影子驾驶的风险特征。
S503可以由风险控制模块416执行,风险控制模块416根据风险识别模块415计算得到的自动驾驶模式的风险特征和影子驾驶模式的风险特征识别自动驾驶系统的潜在缺陷,并更新风险控制策略。
S504可以由安全评估模块422执行,安全评估模块422根据累积的各种驾驶模式的风险特征评估自动驾驶模式的安全性,评估的结果作为安全验证和量产发布的依据。
下面将基于方法500的流程进一步详细阐述本申请提供的自动驾驶安全评估方案。
图6示出了本申请提供的一种自动驾驶安全评估方法。该方法600包括:
S610,获取r b和r c,其中,r b为第一计量单位内车辆的影子驾驶模式的风险值,r c为第一计量单位内车辆基于预设路线的自动驾驶模式的风险值,影子驾驶模式为车辆基于实时路线的自动驾驶模式,所述实时路线为根据处于人工驾驶模式的车辆的位置和运动参数预测的路线,第一计量单位为时间段或路程。
S620,根据r b确定R B,R B为车辆在多个计量单位内的影子驾驶模式的风险值,多个计量单位包括第一计量单位。
S630,根据r c确定R C,R C为车辆在多个计量单位内的基于预设路线的自动驾驶模式的风险值;其中,R B和R C用于确定车辆的自动驾驶模式的安全性是否符合要求。
方法600可以由云端装置420执行,也可以由配置了安全评估模块的车端装置执行。
S610中,r b和r c反映了风险转化为事故的概率,可以是风险转化为事故的概率值,也可以是其它类型的能够量化自动驾驶风险的数值。r b和r c的计量单位可以是时间段,也可以是路程。
可以将风险行为定义为显性的或隐性的违背安全目标的驾驶行为(包括人工驾驶、自动驾驶和影子驾驶),其中,显性表示驾驶行为已转化为事故,隐性表示驾驶行为可能转化为事故。因此,对风险行为的识别包括:
显性风险的识别,例如,碰撞和违反红绿灯规则;
隐性风险的识别,例如,未与前车预留足够的安全距离、即将达到安全行驶速度的最大值仍持续加速、以及十字路口未及时减速;隐性风险还包括异常的行车行为,这些行车行为可能导致他车不能及时反应而造成被动事故,例如,异常的加减速、以及未按道路方向行驶。
通过定义和计算这些风险行为在相应的场景中的特征,可以达到自动识别风险场景的目的。
上述安全目标(safety goal,SG)包括但不限于如下定义:
SG1:避免撞击人或其它物体;
SG2:遵循交通信号灯规则;
SG3:避免冲出道路边界;
SG4:避免非常规行驶(包括超速、异常的加减速以及转向)
针对每一个安全目标,当出现显性风险时,可以定义风险值为1;当无风险时,可以定义风险值为0;隐性风险的风险值为大于0且小于1的数值。下文将详细介绍各个场景中计算风险值的方法。
第一计量单位为任意一个计量单位。基于预设路线的自动驾驶模式即图5所示的自动驾驶模式,该模式为自动驾驶模块413控制车辆的驾驶模式,自动驾驶模块413获取一段路程的起点和终点(即,预设路线)后控制车辆行驶该段路程,司机不参与驾驶。基于实时路线的驾驶模式为司机控制车辆行驶的过程中影子驾驶模块414模拟的自动驾驶模式,影子驾驶模块414根据司机控车时车辆的位置和运动参数不断修正影子轨迹,避免出现司机控车的路线通过A街道,自动驾驶的路线却通过B街道的情况,使得自动驾驶的安全性评估更加客观。
S620和S630中,R B例如可以是多个计量单位的r b的和除以多个计量单位后得到的平均值,也可以是每个计量单位的r b乘以一个加权系数之后的和再除以多个计量单位后得到的平均值;同理,R C例如可以是多个计量单位的r c的和除以多个计量单位后得到的平均值,也可以是每个计量单位的r c乘以一个加权系数之后的和之后再除以多个计量单位得到的平均值。R B和R C也可以是通过其它方法计算得到的值,本申请对计算R B和R C的方法不作限定。
综上,本申请提供的自动驾驶安全评估方法除了收集基于预设路线的自动驾驶模式产生的风险数据外,还收集基于实时路线的自动驾驶模式产生的风险数据,该实时路线为根据处于人工驾驶模式的车辆的位置和运动参数预测的自动驾驶路线(即,影子轨迹),例如,当司机控车改变路线时,影子驾驶模块414可以根据司机控车路线重新规划自动驾驶路线,并基于重新规划的自动驾驶路线确定自动驾驶的安全性,从而能够更加真实地评估自动驾驶的安全性。
可选地,方法600还包括:
S640,计算R B与R C之和。
S641,当R B与R C之和小于或等于自动驾驶风险阈值时,确定车辆的自动驾驶模式的安全性符合要求。
上述实施例可以由确定R B和R C的装置实施,无需将R B和R C发送至其它设备以及通 过其它设备评估车辆的自动驾驶模式的安全性,从而可以提高自动驾驶模式安全评估的实时性。
S641中,自动驾驶风险阈值表示每个计量单位内自动驾驶模式的风险值的阈值。自动驾驶风险阈值可以是依据统计结果设定的值,例如,车辆行驶的一百万公里的里程中平均每公里的自动驾驶风险值的最大值,若车辆行驶一百万公里后R B+R C的值小于或等于该最大值,则车辆的自动驾驶功能符合安全性要求,可以大批量生产;若车辆行驶一百万公里后R B+R C的值大于该最大值,则车辆的自动驾驶功能不符合安全性要求,不能大批量生产,需要进行改进。
还可以设定每百万公里的R B+R C中R B和R C分别对应的里程,例如,可以设定每百万公里的R B+R C中R C对应的里程不少于10万公里,从而可以更加客观真实地评估自动驾驶的安全性。
作为一个可选的实施方式,所述自动驾驶风险阈值为a(R A+R′ A),其中,a为风险容忍系数,R A例如是车辆在多个计量单位内的人工驾驶的模式的风险值的平均值,车辆在多个计量单位内通过的路段为第一路段,R′ A例如是车辆在多个计量单位之外基于人工驾驶模式通过第一路段的风险值的平均值。
可以将自动驾驶的风险值与人工驾驶的风险值做对比,并设定自动驾驶的风险值能够被接受的容忍系数,从而可以判断自动驾驶的安全性是否满足要求。
例如,R A为人工驾驶模式下车辆行驶五十万公里后得到的风险值,该五十万公里通过的路段为第一路段,R′ A为人工驾驶模式下车辆再次通过该第一路段后得到的风险值;R B+R C为自动驾驶行驶一百万公里后得到的风险值,a表示行驶相同长度的路程的情况下,自动驾驶的风险值需要小于人工驾驶的风险值与a的乘积,才可以认定自动驾驶模式的安全性符合要求,a通常为大于0且小于1的常数。
作为一个可选的实施方式,根据r b确定R B,以及根据r c确定R C,包括:
S650,根据
Figure PCTCN2019072074-appb-000010
确定R B或R C,其中,event表示风险场景,s event表示该风险场景中事故的严重程度的系数,X表示所述多个计量单位,当r i表示r b时,R i表示R B,当r i表示r c时,R i表示R C
s event反映了不同的风险场景中事故的严重程度,r i反映了风险转化为事故的概率,∑则反映了多个计量单位的风险的累计。
例如,在碰撞场景中,撞人的严重程度远高于撞其它物体的严重程度,因此,可以将撞人的s event设置为1,将撞其它物体的s event可以根据物体的性质从0.1-0.9中选择。
又例如,违反交通规则、偏离道路范围或方向行驶、超速、违反驾驶常规等行为是构成事故的几个主要因素,因此,在一种方案中,s event可以依据事故统计数据进行定义,当车辆出现上述几种行为中的任意一种行为时,相应的s event等于0.01。
X可以是与R i对应的车辆行驶的里程或时间,R i则反映了平均每个计量单位(例如,每公里或每小时)内风险转化为事故的概率。上述公式可以客观地反映自动驾驶的风险。
作为一个可选的实施方式,获取r b和r c,包括:
S611,当风险场景为碰撞场景时,根据下列参数中的至少一个确定r b和r c:车辆相对于碰撞物的速度Δv、基于碰撞物的运动轨迹预估的车辆与碰撞物发生碰撞所需的时间 TTC、基于车辆与碰撞物之间的距离预估的车辆的刹车时间THW、以及车辆与碰撞物之间的距离,其中,Δv与r b或r c正相关,TTC与r b或r c负相关,THW与r b或r c负相关,车辆与碰撞物之间的距离与r b或r c负相关。
在碰撞场景中,相对运动速度越大,风险转化为事故的概率越大,即,Δv与r b或r c正相关;距离碰撞的时间越长,车辆调整行驶路线避免碰撞的可能性越大,碰撞发生的概率越小,因此,TTC与r b或r c负相关;刹车可用时间越长,碰撞发生的概率越小,因此,THW与r b或r c负相关;车辆与碰撞物的距离越远,车辆调整行驶路线避免碰撞的可能性越大,碰撞发生的概率越小,因此,车辆与所述碰撞物之间的距离与r b或r c负相关。
图7示出了本申请提供的一种碰撞场景中的风险示意图。
图7中自车的轨迹是从状态(t2,s2,v2,a2)到状态(t0,s0,v0,a0)之间的线段,他车的轨迹是从状态(t1,s1,v1,a1)到状态(t’0,s’0,v’0,a’0)之间的线段,两车相向而行,在图中所示的两条线的交叉点相撞。
上述两条轨迹均为自车预测的轨迹,自车根据传感器的感知结果识别未来一段时间(例如5秒)内与自车轨迹有交集的对象,未来一段时间内的碰撞风险值取自车与上述对象的之间的碰撞风险值中的最大值。
例如,可以根据r SG1=f(t) max确定r b和r c,r SG1为碰撞场景中的r b或r c,其中,t∈{t event,h(t event,obj)=1},f(t)=Δv(t)·exp(-min{r 0TTC(t),r 1THW(t),r 2s}),{t event,h(t event,obj)=1}表示发生碰撞风险的时间段,TTC(t)表示t时刻的TTC,THW(t)表示t时刻的THW,
Figure PCTCN2019072074-appb-000011
r 0、r 1和r 2分别为归一化参数,v为t时刻车辆的速度,Δv(t)为t时刻的Δv,exp为以自然常数为底的指数,s表示车辆与碰撞物之间的距离。
上述公式中,t∈{t event,h(t event,obj)=1}表示已识别的产生碰撞风险的时间段,h(t event,obj)表示碰撞风险事件(obj)在一段时间(t event)内引发了潜在的冲突事件,或者可以理解为,h(t event,obj)表示碰撞风险事件(obj)的出现导致在一段时间(t event)内发生碰撞的概率增加,h(t event,obj)=1表示发生了碰撞风险事件(obj),h(t event,obj)=0表示未发生碰撞风险事件。由于风险事件从出现到消失通常会持续一段时间,因此,当车端装置410或云端装置420识别到碰撞风险事件发生(即,h(t event,obj)=1)时,可以选取识别到该风险事件的时刻附近的一段时间t event,计算该段时间内的风险值,该段时间包括识别到该碰撞风险事件的时刻。各个归一化参数用于表示各个参数的权重,例如,各个归一化参数可以取大于或等于0且小于或等于1的值。
例如,当两辆车的相对速度大于40km/h时,如果发生碰撞,有极大概率造成人员伤亡;若两辆车预期碰撞时间(即TTC)小于1s,则有极大概率无法避免辆车相撞;当THW小于1s时,若他车制动,则自车有极大概率无法避免辆车碰撞。
上述“相对速度大于40km/h”、“TTC小于1s”和“THW小于1s”均为碰撞风险事件(obj),可以预测碰撞发生时刻,并在预测的碰撞发生时刻附近选取一段时间(t event)计算该段时间内的碰撞风险值(发生碰撞的概率),取该段时间内多个碰撞风险值中最大 的值(即,f(t) max)作为该段时间的碰撞风险值。
作为一个可选的实施方式,获取r b和r c,包括:
S612,当风险场景为交通信号灯场景时,根据车辆的速度以及车辆到交通信号灯的距离确定r b和r c,其中,车辆的速度与r b或r c正相关,车辆到交通信号灯的距离与r b或r c负相关。
在交通信号灯场景中,车辆的速度越高,违反交通信号灯(例如闯红灯)的概率越大,因此,车辆的速度与r b或r c正相关;车辆到交通信号灯的距离越远,用于车辆改变行驶状态的时间越长,违反交通信号灯的概率越小,因此,车辆到交通信号灯的距离与r b或r c负相关。
图8示出了本申请提供的一种交通信号灯场景中的风险示意图。
交通信号灯场景的风险主要是违反交通信号灯所指示的规则,相应的风险值可以根据自车的速度得到。如图8所示,自车距离交通信号灯50m,车辆距离路口停止线的距离s、自车的速度v以及车辆的加速度a构成当前车辆状态。右图为自车的距离-速度曲线,当交通信号灯为禁止通行信号时,自车需减速,在停止线前将速度降为0。不同的加速度形成不同的速度-距离曲线。
当预计采用普通加速度即可在停止线前停下时,则确定自车当前的状态无违反交通信号灯的风险,其中,普通加速度指的是绝对值小于预设加速度的绝对值的加速度,例如,预设加速度的值为0.4g,若自车采用-0.3g的加速度即可在停止线前停下,则自车的r b或r c等于0。
当预计采用最大加速度仍不能在停止线前停下时,则确定自车当前的状态必然导致违反交通信号灯的事故,即,此时的风险为显性风险,自车的r b或r c等于1。
当自车的数据-距离曲线位于图8所示的曲线1和曲线2之间时,自车存在违反交通信号灯的隐性风险,此时,自车在停止线前将速度降为0所需的加速度的绝对值越大,r b或r c的取值越大。
例如,可以根据r SG2=f(t) max确定r b和r c,r SG2为交通信号灯场景中的r b或r c,其中,t∈{t event,h(t event,Light)=1},
Figure PCTCN2019072074-appb-000012
r 3为归一化参数,v(t)表示车辆在t时刻的速度,g为重力加速度常数,s(t)表示t时刻车辆到交通信号灯的距离。
上述公式中,t∈{t event,h(t event,Light)=1}表示已识别的产生交通信号灯场景风险的时间段,h(t event,Light)表示交通信号灯风险事件(Light)的出现导致在一段时间(t event)内发生违反交通信号灯的概率增加,h(t event,Light)=1表示发生了交通信号灯风险事件(Light),h(t event,Light)=0表示未发生交通信号灯风险事件。由于风险事件从出现到消失通常会持续一段时间,因此,当车端装置410或云端装置420识别到交通信号灯场景的风险事件发生(即,h(t event,Light)=1)时,可以选取识别到该风险事件的时刻附近的一 段时间t event,计算该段时间内的风险值,该段时间包括识别到该碰撞风险事件的时刻。归一化参数用于表示参数的权重,例如,各个归一化参数可以取大于或等于0且小于或等于1的值。
例如,车辆距离路口50m,该路口的交通信号灯显示禁行信号,该车辆的速度为100km/h,则有极大的概率发生违法交通信号灯(例如闯红灯)的风险,上述“车辆距离路口50m,该路口的交通信号灯显示禁行信号,该车辆的速度为100km/h”即为交通信号灯风险事件(Light),可以预测闯红灯时刻,并在预测的闯红灯时刻附近选取一段时间(t event)计算该段时间内的闯红灯风险值(发生闯红灯的概率),取该段时间内多个闯红灯风险值中最大的值(即,f(t) max)作为该段时间的闯红灯风险值。
作为一个可选的实施方式,获取r b和r c,包括:
S613,当风险场景为冲出道路边界场景时,根据预估的车辆冲出车道时间确定r b和r c,其中,预估的车辆冲出车道时间与r b或r c负相关。
在冲出道路边界场景中,预估的车辆冲出车道的时间越长,用于车辆调整行驶状态的时间越长,车辆冲出道路的概率也就越小,因此,预估的车辆冲出车道时间与r b或r c负相关。
例如,可以根据
Figure PCTCN2019072074-appb-000013
确定r b和r c,r SG3为冲出道路边界场景中的r b或r c,其中,TLC threshold表示车辆冲出道路所需时间的临界值,
Figure PCTCN2019072074-appb-000014
e为自然常数,t lc为预估的车辆冲出车道时间,y为车辆在道路边界的横向偏移量,W road为道路的宽度,v long为车辆的纵向速度,v lat为车辆的横向速度,φ-φ d表示车辆的航向与道路切线方向的夹角,r 4为归一化参数,归一化参数用于表示参数的权重。
作为一个可选的实施方式,获取r b和r c,包括:
S614,当风险场景为非常规行驶场景(例如,车辆当前具有横向加速度的行驶场景)时,根据下列参数中的至少一个确定r b和r c:车辆的纵向加速度、车辆的横向加速度以及车辆的速度,其中,车辆的纵向加速度与r b或r c正相关,车辆的横向加速度与r b或r c正相关,车辆的速度与r b或r c正相关。上述纵向加速度指的是加速度方向与速度方向相同或相反的加速度(例如,直线行驶时踩油门或踩刹车获得的加速度),上述横向加速度指的是加速度方向与速度方向具有不为0的夹角的加速度(例如,转弯时导致车辆发生漂移的加速度)。
纵向加速度越大,车辆超速或者急刹车的概率越大,因此,车辆的纵向加速度与r b或r c正相关;横向加速度越大,车辆急转弯的概率越大,因此,车辆的横向加速度与r b或r c正相关;车辆的速度越大,出现超速等非常规行驶的概率越大,因此,车辆的速度与r b或r c正相关。
例如,可以根据r SG4=r 5确定r b和r c,r SG4为非常规行驶场景中的r b或r c,其中,(|a long|>a threshold1)||(|a lat|>a threshold2)||(v≥v safe),|a long|表示所述车辆的纵向加速度的绝对值, |a lat|表示所述车辆的横向加速度的绝对值,a threshold1表示安全行驶状态下横向加速度的阈值,例如可以为0.5g,a threshold2表示安全行驶状态下纵向加速度的阈值,例如可以为0.7g,v表示所述车辆的速度,v safe表示安全行驶状态下的速度阈值,可以参考道路限速确定,r 5为归一化参数。
上述公式中,(|a long|>a threshold1)||(|a lat|>a threshold2)||(v≥v safe)表示车辆的横向加速度、纵向加速度和速度中的至少一个超过了相应的阈值。
上文详细介绍了本申请提供的自动驾驶安全评估方法,若云端装置420从车端装置410接收的信息为r b或r c,则云端装置420无需再执行S611~S614,若云端装置420从车端装置410接收的信息为传感器感知的数据,则云端装置420还需要执行S611~S614以获取r b和r c
应理解,上述实施例仅描述了生成r b或r c的方案,S611~S614同样适用于计算人工驾驶的风险值r a
由方法600可知,影响自动驾驶安全评估方法的技术效果的一个重要因素是风险值是否能够反映真实的风险水平,下面,将详细描述本申请提供的确定车辆的驾驶风险的方法。
图9示出了本申请提供的一种确定车辆的驾驶风险的方法。该方法900应用于具有自动驾驶模式和人工驾驶模式的车辆,方法900包括:
S910,确定一个时间段内车辆的驾驶模式和风险场景,该时间段的起始点为第一时刻,该时间段的终点为第二时刻,所述驾驶模式包括自动驾驶模式和/或人工驾驶模式。
S920,在当前时刻根据风险场景对应的风险识别方式确定所述时间段内驾驶模式的风险值,其中,当前时刻位于第一时刻之后,且当前时刻位于第二时刻之前。
方法900例如可以由车端装置410执行,车端装置410在车辆行驶一段路程所需的多个时间段中选取一个时间段,并在该时间段内的一个时刻(即,当前时刻)利用已知的行驶信息和预测的行驶信息确定该时间段内相应的驾驶模式的风险值,例如,确定的时间段为1秒,当前时刻为0.4秒,若已过去的0.4秒内出现了超速行驶的状况,则车辆可以确定该时间段内的驾驶模式的非常规驾驶风险值为1,即,发生超速的概率为100%;若已过去的0.4秒内未出现超速行驶的状况,且车辆根据当前车辆的速度以及加速度确定未来0.6秒内车速将保持在法定行驶速度内,则车辆可以确定该时间段内的非常规驾驶风险值为一个小于1的数值,即,发生超速的概率小于100%。本实施例将行驶一段路程所需的时间段分割为多个时间段,并根据每个时间段中的交通上下文(即,在当前时刻之前和在当前时候之后的与交通环境相关的信息)确定每个时间段内的驾驶模式的风险值,从而能够更加客观地识别风险。
上述确定一个时间段的方式仅是举例说明,车端装置410还可以通过其它方法确定一个时间段,例如,根据路程和速度确定行驶一段距离所需的时间,该段时间即为S910所述的“一个时间段”;再例如,在当前时刻确定第一时刻和第二时刻,从而确定了S910所述的“一个时间段”。
上述实施例可以确定自动驾驶模式和/或人工驾驶模式的风险值,其中,自动驾驶模式包括影子驾驶模式,基于影子驾驶模式的风险值评估自动驾驶系统的安全性是本申请提供的自动驾驶安全评估方法能够取得相比于现有技术更好的技术效果的一个重要因素。
影子驾驶模式的目的是在人工驾驶模式下,通过系统后台运行的软件和硬件生成模拟的自动驾驶控车轨迹,识别自动驾驶系统潜在的风险,从而进一步识别自动驾驶系统的隐患,并累计安全评估所需数据。
要达到以上目的,一大难点在于车辆在人工驾驶模式下,自动驾驶模块并不能真正控制车辆,影子轨迹如何逼近自动驾驶系统的真实风险水平。
直接基于传感器的感知数据生成的自动驾驶控车轨迹无法代表自动驾驶系统的真实风险水平,会造成自动驾驶安全评估结果失真。
图10示出了一种自动驾驶控车轨迹的示意图。虚线箭头表示人工驾驶模式下车辆的实际轨迹,实线箭头表示根据传感器的感知数据模拟的自动驾驶轨迹,由于感知数据指示左侧车道有车,因此自动驾驶系统会模拟出继续沿原车道前行的轨迹,实际上,车辆并未沿原车道行驶,而是进行了变道,上述根据传感器的感知数据模拟的自动驾驶轨迹不能反映真实的驾驶情况,导致基于轨迹的自动驾驶安全评估结果失真。
下面将详细介绍本申请提供的影子驾驶模式如何解决上述结果失真的问题。
作为一个可选的实施方式,在S920之前,方法900还包括:
S901,确定车辆在第一时刻的位置和运动参数,其中,车辆在第一时刻基于人工驾驶模式行驶。
S902,根据车辆在第一时刻的位置和运动参数预测车辆在第三时刻的位置,第三时刻晚于第一时刻。
S903,根据第一时刻的位置和第三时刻的位置预测车辆从第一时刻的位置到第三时刻的位置的影子驾驶模式,影子驾驶模式为基于车辆处于人工驾驶模式下的位置和运动参数预测的车辆的自动驾驶模式。
其中,S920包括:
S921,在当前时刻根据风险场景对应的风险识别方式确定r a、r b和r c,其中,r a为所述时间段内人工驾驶模式的风险,r b为所述时间段内影子驾驶模式的风险,r c为所述时间段内基于预设路线的自动驾驶模式的风险。
车辆可以根据该车辆在第一时刻的位置和运动参数预测未来一段时间后车辆的位置,即,预测车辆在第三时刻的位置,其中,车辆在第一时刻处于人工驾驶状态,随后,车辆预测该车辆从第一时刻的位置到第三时刻的位置的影子驾驶模式。上述实施例根据人工驾驶模式下的位置和运动参数不断修正自动驾驶的路线,使得自动驾驶模式的路线更加接近人工驾驶模式的路线,从而可以获取更加客观的自动驾驶风险数据。
上述第三时刻可以是S920所述的当前时刻,也可以是第二时刻,还可以是第二时刻之后的一个时刻。
图11示出了本申请提供的一种影子驾驶模式的流程图。
在司机驾驶汽车时,影子驾驶模块414在第一时刻根据车辆的运动参数(例如,速度和加速度)识别司机的驾驶意图,即,S901~S903所述的内容,随后将识别到的司机驾驶意图作为影子驾驶模式414的输入信息生成影子轨迹,从而提高了影子轨迹与司机控车轨迹的重合度。
上述司机的驾驶意图包括但不限于:沿当前车道行驶、转弯和换道中的至少一种。
图12示出了本申请提供的一种识别司机意图的方法的流程图。
从T时刻(相当于第一时刻)开始,至T+1时刻(相当于第三时刻)为止,若自车轨迹完整跨越车道线,则判定司机的驾驶意图为换道。
从T时刻开始,至T+1时刻为止,若自车轨迹沿左拐或右拐车道,完成转弯轨迹;或者,在T时刻之前与T时刻相邻的一段时间内偏航角速度(YawRate)持续大于角速度阈值,则判定司机的驾驶意图为转弯。
从T时刻开始,至T+1时刻为止,若自车轨迹未完整跨越当前车道,则判定司机的驾驶意图为沿当前车道行驶,其中,道内避障、让行前车以及跟随前车等行为均属于沿当前车道行驶。
上述实施例利用当前信息预测未来时刻的信息,并利用预测的未来时刻的信息判定当前司机的驾驶意图,由此消除了根据传感器的感知数据判定司机意图导致的偏差。
图13示出了本申请提供的基于影子驾驶模式预测影子轨迹的示意图。
图13所示的影子轨迹与司机控车轨迹大致相同,但略有差异,根据司机控车轨迹与影子轨迹分别得到不同的风险值,通过该风险值的差异可评估人工驾驶模式与自动驾驶模式哪个更加安全。
例如,当司机驾驶意图为风险意图(例如,侵略性的切入)时,若此时影子驾驶模块414正确识别到风险,则不产生符合司机驾驶意图的影子轨迹,而选择更安全的决策(例如,保持车道内顺行)生成影子轨迹。通过对影子轨迹与真实控车结果的风险识别,可以得到司机的危险意图场景,以及此时影子轨迹做出的更安全的决策(或相反)。
在较短的时间片段内,例如3s内,由于司机驾驶和影子轨迹大致相同,可近似为自动驾驶系统闭环控制,此时影子轨迹造成的风险场景,与真实系统控车时可能造成的风险场景类似。图11所示的计算影子轨迹步骤计算出的影子轨迹风险值的示意图如图14所示。
图14中,实线表示按照S611~S614计算的司机控车轨迹得到的风险值,虚线为按照S611~S614计算的每一个时间片段(例如,第一时刻至第三时刻的时间段)对应的影子轨迹的风险值。
示例1中,司机控车轨迹未产生风险(r a等于0),而影子轨迹产生了新的风险(r b大于0),导致风险指标从安全区域(r=0)落入风险区域(0<r<1),由此可识别出引起风险的自动驾驶系统的隐患,即,识别出图2的F区所示的“未知隐患”(斜线部分)。
示例2中,司机控车轨迹持续落入风险区域,而影子轨迹从风险区域回到安全区域,说明通过系统的模拟控车,降低了行车风险,该数据可作为自动驾驶系统安全性能优于司机控车的一个证据。
示例3中,司机控车轨迹和影子控车轨迹均持续落入风险区域,但影子轨迹的风险值小于司机控车,说明通过自动驾驶系统的模拟控车降低了驾驶风险,该数据可作为自动驾驶系统安全性能优于司机控车的一个证据。
上述三个示例中的数据均可作为方法600进行自动驾驶安全评估所使用的数据。
由图12至图14所示的内容可知,由于影子轨迹生成方法与自动驾驶控车算法一致;影子轨迹的风险计算方法和自动驾驶/司机驾驶轨迹风险计算方法一致,从而可以保证影子轨迹的风险代表自动驾驶系统的风险。
在评估自动驾驶系统的安全性时,作为一个可选的实施方式,方法900还包括:
S930,向安全中心发送r a、r b和r c,r a、r b和r c用于评估人工驾驶模式和影子驾驶模 式的安全性。
车辆获取r a、r b或r c后可以向安全中心发送上述风险数据,用于安全中心评估人工驾驶模式和影子驾驶模式的安全性,减小了车载处理器的工作负担。上述安全中心例如是云端装置420。
对于自动驾驶汽车来说,评估其自动驾驶模式的安全性是一个长期的过程,即使自动驾驶模式的安全性评估通过,也不意味着自动驾驶模式百分之百安全,因此,有必要设计一种即时控制风险的方案,即,通过风险控制模块416减小或者消除风险。
图15示出了本申请提供的一种风险控制的方法的示意性流程图。
如图15所示,风险场景数据包括但不限于传感器411获取的感知数据、自动驾驶模块413生成的轨迹和影子驾驶模块414生成的轨迹中的至少一个,风险识别模块415根据上述风险场景数据确认当前场景是否为风险场景,若风险识别模块415确认当前场景为非风险场景,且事后发现该当前场景实际为风险场景,则可以调整风险识别机制,例如,调整风险识别阈值参数,修改风险识别算法等。若风险识别模块415确认当前场景为风险场景,且判断该风险场景是否为自动驾驶系统的隐患导致,则风险识别模块415向风险策略模块421上报风险场景,风险策略模块421向风险控制模块416发送策略更新信息,更新车端装置410的风险控制策略,将风险场景排除在安全场景围栏之外。
作为一个可选的实施方式,方法900还包括:
S940,接收安全中心发送的安全场景信息,所述安全场景信息用于指示自动驾驶模式的安全驾驶交通场景集合。
S950,根据安全场景信息确定自动驾驶模式的安全驾驶交通场景集合。
S960,若当前交通场景不属于安全驾驶交通场景集合,车辆播放和/或显示提示司机进行人工驾驶的提示信息,和/或,车辆执行安全停靠处理。
安全中心可以根据车辆上报的风险数据识别风险场景,通过向车辆发送安全场景信息更新车辆的风险识别机制和风险控制策略,例如,若风险是由自动驾驶系统的隐患导致的,则安全场景信息可以指示车辆将风险场景排除在安全驾驶交通场景集合之外,车辆接收到安全场景信息后更新安全驾驶交通场景集合,若当前交通场景不属于更新后的安全驾驶交通场景集合,则车辆采取安全措施,例如,播放和/或显示提示司机进行人工驾驶的提示信息,和/或,执行安全停靠处理。
上述安全驾驶交通场景集合也可称为安全驾驶场景围栏,由环境、静态场景、动态场景、自车行为构成。在一种技术方案中,场景围栏的定义如图16所示。
图16中,圆括号内的内容为自动驾驶系统的安全驾驶场景围栏所包含的内容。云端装置420根据识别的自动驾驶系统隐患及时发送安全场景信息,更新安全驾驶场景围栏的内容。在车端装置410,风险控制模块416根据地图输入、实时感知结果获得场景围栏各要素的识别结果,并根据云端装置420发送的安全场景信息更新安全驾驶场景围栏的围内容,在自动驾驶模块413控车行驶过程中,若风险控制模块416确定当前场景超出安全驾驶场景围栏范围,则提醒司机接管,或将车辆控制至安全地带停车。
在图16所示的一个示例(方括号中的内容)中,一个风险事件是由于路口宽度大于20m,导致系统无法准确识别某一类型的红绿灯,造成潜在风险,则在更新的安全驾驶场景围栏中,路口宽度小于20m。当车辆行驶再次行驶至路口宽度大于或等于20m的路口 时,则自动驾驶系统采取保守的控制策略(例如,提示司机或靠边停车)。
在图16所示的另一个示例(方括号中的内容)中,一个风险事件是由于自动驾驶系统无法及时测量低速车辆的速度,造成潜在风险,则在更新的安全驾驶场景围栏中,动态场景的车流速度适用值为10~50千米每小时(kph)。当自动驾驶系统检测到当前车流小于10kph时,则自动驾驶系统提示司机接管汽车,或平缓靠边停车。
此外,还可调整司机在环策略,例如调整提醒间隔等。
上文详细介绍了本申请提供的评估自动驾驶安全性和确定车辆的驾驶风险的方法的示例。可以理解的是,车端装置和云端装置为了实现上述功能,其包含了执行各个功能相应的硬件结构和/或软件模块。本领域技术人员应该很容易意识到,结合本文中所公开的实施例描述的各示例的单元及算法步骤,本申请能够以硬件或硬件和计算机软件的结合形式来实现。某个功能究竟以硬件还是计算机软件驱动硬件的方式来执行,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。
本申请可以根据上述方法示例对车端装置和云端装置进行功能单元的划分,例如,可以对应各个功能划分各个功能单元,也可以将两个或两个以上的功能集成在一个处理单元中。上述集成的单元既可以采用硬件的形式实现,也可以采用软件功能单元的形式实现。需要说明的是,本申请中对单元的划分是示意性的,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。
在采用集成的单元的情况下,图17示出了上述实施例中所涉及的车端装置的一种可能的结构示意图。车端装置1700包括:处理单元1702和通信单元1703。处理单元1702用于对车端装置1700的动作进行控制管理,例如,处理单元1702用于支持车端装置1700执行S620和/或用于本文所描述的技术的其它过程。通信单元1703用于支持车端装置1700与云端装置的通信。车端装置1700还可以包括存储单元1701,用于存储车端装置1700的程序代码和数据。
在一个可选的实施例中,处理单元1702用于执行S610~S630,其中,当处理单元1702执行S610时,处理单元1702也可称为获取单元1702。存储单元1701需要存储处理单元1702在执行上述步骤的过程中产生的中间数据,中间数据例如是r b和r c,以便于处理单元1702根据多个计量单位内的影子驾驶模式的风险值和基于预设路线的自动驾驶模式的风险值确定车辆的自动驾驶模式的安全性是否符合要求。
在另一个可选的实施例中,处理单元1702用于执行S610,并控制通信单元1703向云端设备发送r b和r c,以便于云端设备执行S620~S630。
在第三个可选的实施例中,处理单元1702用于控制通信单元1703向云端设备发送用于计算r b和r c的原始数据(例如,传感器数据),以便于云端设备执行S610~S630。
在第四个可选的实施例中,处理单元1702用于执行S910和S920,可选地,处理单元1702用于控制通信单元1703执行S930。
处理单元1702可以是处理器或控制器,例如可以是中央处理器(central processing unit,CPU),通用处理器,数字信号处理器(digital signal processor,DSP),专用集成电路(application-specific integrated circuit,ASIC),现场可编程门阵列(field programmable gate array,FPGA)或者其他可编程逻辑器件、晶体管逻辑器件、硬件部件或者其任意组 合。其可以实现或执行结合本申请公开内容所描述的各种示例性的逻辑方框,模块和电路。所述处理器也可以是实现计算功能的组合,例如包含一个或多个微处理器组合,DSP和微处理器的组合等等。通信单元1703可以是收发器等。存储单元1701可以是存储器。
当处理单元1702为处理器,通信单元1703为收发器,存储单元1701为存储器时,本申请所涉及的车端装置可以为图18所示的车端装置1800。
参阅图18所示,该车端装置1800包括:处理器1802、收发器1803、存储器1801。其中,收发器1803、处理器1802以及存储器1801可以通过内部连接通路相互通信,传递控制和/或数据信号。
本领域的技术人员可以清楚地了解到,为了描述的方便和简洁,上述描述的装置和单元的具体工作过程,可以参考前述方法实施例中的对应过程,在此不再赘述。
在采用集成的单元的情况下,图19示出了上述实施例中所涉及的云端装置的一种可能的结构示意图。云端装置1900包括:处理单元1902和通信单元1903。处理单元1902用于对云端装置1900的动作进行控制管理,例如,处理单元1902用于支持云端装置1900执行S620和/或用于本文所描述的技术的其它过程。通信单元1903用于支持云端装置1900与车端装置的通信。云端装置1900还可以包括存储单元1901,用于存储云端装置1900的程序代码和数据。
在一个可选的实施例中,处理单元1902用于控制通信单元1903执行S610,即,从车端装置接收r b和r c,此时,通信单元1903也可称为获取单元1903,处理单元1902还用于执行S620~S641。存储单元1901需要存储处理单元1902在执行上述步骤的过程中产生的中间数据,中间数据例如是r b和r c,以便于处理单元1902根据多个计量单位内的影子驾驶模式的风险值和基于预设路线的自动驾驶模式的风险值确定车辆的自动驾驶模式的安全性是否符合要求。
在另一个可选的实施例中,处理单元1902用于控制通信单元1903从车端装置接收用于计算r b和r c的原始数据(例如,传感器数据),处理单元1902还用于根据通信单元1903接收的原始数据得到r b和r c,此时,处理单元1902也可以称为获取单元1902,处理单元1902还用于执行S620~S641,存储单元1901需要存储处理单元1902在执行上述步骤的过程中产生的中间数据,中间数据例如是r b和r c,以便于处理单元1902根据多个计量单位内的影子驾驶模式的风险值和基于预设路线的自动驾驶模式的风险值确定车辆的自动驾驶模式的安全性是否符合要求。
在第三个可选的实施例中,处理单元1902用于执行S910和S920,可选地,处理单元1902用于控制通信单元1903执行S930。
处理单元1902可以是处理器或控制器,例如可以是CPU,通用处理器,DSP,ASIC,FPGA或者其他可编程逻辑器件、晶体管逻辑器件、硬件部件或者其任意组合。其可以实现或执行结合本申请公开内容所描述的各种示例性的逻辑方框,模块和电路。所述处理器也可以是实现计算功能的组合,例如包含一个或多个微处理器组合,DSP和微处理器的组合等等。通信单元1903可以是收发器等。存储单元1901可以是存储器。
当处理单元1902为处理器,通信单元1903为收发器,存储单元1901为存储器时,本申请所涉及的云端装置可以为图20所示的云端装置2000。
参阅图20所示,该云端装置2000包括:处理器2002、收发器2003、存储器2001。 其中,收发器2003、处理器2002以及存储器2001可以通过内部连接通路相互通信,传递控制和/或数据信号。
本领域的技术人员可以清楚地了解到,为了描述的方便和简洁,上述描述的装置和单元的具体工作过程,可以参考前述方法实施例中的对应过程,在此不再赘述。
本申请还提供一种评估自动驾驶安全性的系统,其包括前述的一个或多个车端装置,和,一个或多个云端装置。
应理解,本申请中的处理器可以是一种集成电路芯片,具有信号的处理能力。在实现过程中,上述方法实施例的各步骤可以通过处理器中的硬件的集成逻辑电路或者软件形式的指令完成。上述的处理器可以是通用处理器、数字信号处理器(digital signal processor,DSP)、专用集成电路(application specific integrated circuit,ASIC)、现场可编程门阵列(field programmable gate array,FPGA)或者其它可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件。可以实现或者执行本申请实施例中的公开的各方法、步骤及逻辑框图。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。结合本申请各实施例所公开的方法的步骤可以直接体现为硬件译码处理器执行完成,或者用译码处理器中的硬件及软件模块组合执行完成。软件模块可以位于随机存储器,闪存、只读存储器,可编程只读存储器或者电可擦写可编程存储器、寄存器等本领域成熟的存储介质中。该存储介质位于存储器,处理器读取存储器中的信息,结合其硬件完成上述方法的步骤。
可以理解,本申请实施例中的存储器可以是易失性存储器或非易失性存储器,或可包括易失性和非易失性存储器两者。其中,非易失性存储器可以是只读存储器(read-only memory,ROM)、可编程只读存储器(programmable ROM,PROM)、可擦除可编程只读存储器(erasable PROM,EPROM)、电可擦除可编程只读存储器(electrically EPROM,EEPROM)或闪存。易失性存储器可以是随机存取存储器(random access memory,RAM),其用作外部高速缓存。通过示例性但不是限制性说明,许多形式的RAM可用,例如静态随机存取存储器(static RAM,SRAM)、动态随机存取存储器(dynamic RAM,DRAM)、同步动态随机存取存储器(synchronous DRAM,SDRAM)、双倍数据速率同步动态随机存取存储器(double data rate SDRAM,DDR SDRAM)、增强型同步动态随机存取存储器(enhanced SDRAM,ESDRAM)、同步连接动态随机存取存储器(synchlink DRAM,SLDRAM)和直接内存总线随机存取存储器(direct rambus RAM,DR RAM)。应注意,本文描述的系统和方法的存储器旨在包括但不限于这些和任意其它适合类型的存储器。
本申请还提供了一种计算机可读介质,其上存储有计算机程序,该计算机程序被计算机执行时实现上述任一方法实施例的功能。
本申请还提供了一种计算机程序产品,该计算机程序产品被计算机执行时实现上述任一方法实施例的功能。
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(digital subscriber line,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,高密度数字视频光盘(digital video disc,DVD))、或者半导体介质(例如,固态硬盘(solid state disk,SSD))等。
本申请还提供了一种处理装置,包括处理器和接口;所述处理器,用于执行上述任一方法实施例所描述的步骤。
应理解,上述处理装置可以是一个芯片,所述处理器可以通过硬件来实现也可以通过软件来实现,当通过硬件实现时,该处理器可以是逻辑电路、集成电路等;当通过软件来实现时,该处理器可以是一个通用处理器,通过读取存储器中存储的软件代码来实现,改存储器可以集成在处理器中,可以位于所述处理器之外,独立存在。
应理解,说明书通篇中提到的“实施例”意味着与实施例有关的特定特征、结构或特性包括在本申请的至少一个实施例中。因此,在整个说明书各个实施例未必一定指相同的实施例。此外,这些特定的特征、结构或特性可以任意适合的方式结合在一个或多个实施例中。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。
另外,本文中术语“系统”和“网络”在本文中常被可互换使用。本文中术语“和/或”,仅仅是一种描述关联对象的关联关系,表示可以存在三种关系,例如,A和/或B,可以表示:单独存在A,同时存在A和B,单独存在B这三种情况。另外,本文中字符“/”,一般表示前后关联对象是一种“或”的关系。
应理解,在本申请实施例中,“与A相应的B”表示B与A相关联,根据A可以确定B。但还应理解,根据A确定B并不意味着仅仅根据A确定B,还可以根据A和/或其它信息确定B。
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示例的单元及算法步骤,能够以电子硬件、计算机软件或者二者的结合来实现,为了清楚地说明硬件和软件的可互换性,在上述说明中已经按照功能一般性地描述了各示例的组成及步骤。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。
所属领域的技术人员可以清楚地了解到,为了描述的方便和简洁,上述描述的系统、装置和单元的具体工作过程,可以参考前述方法实施例中的对应过程,在此不再赘述。
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,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。另外,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,也可以是电的,机械的或其它的形式连接。
作为分离部件说明的单元可以是或者也可以不是物理上分开的,作为单元显示的部件可以是或者也可以不是物理单元,即可以位于一个地方,或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部单元来实现本申请实施例方案的目的。
另外,在本申请各个实施例中的各功能单元可以集成在一个处理单元中,也可以是各个单元单独物理存在,也可以是两个或两个以上单元集成在一个单元中。上述集成的单元既可以采用硬件的形式实现,也可以采用软件功能单元的形式实现。
通过以上的实施方式的描述,所属领域的技术人员可以清楚地了解到本申请可以用硬件实现,或固件实现,或它们的组合方式来实现。当使用软件实现时,可以将上述功能存储在计算机可读介质中或作为计算机可读介质上的一个或多个指令或代码进行传输。计算机可读介质包括计算机存储介质和通信介质,其中通信介质包括便于从一个地方向另一个地方传送计算机程序的任何介质。存储介质可以是计算机能够存取的任何可用介质。以此为例但不限于:计算机可读介质可以包括RAM、ROM、EEPROM、CD-ROM或其他光盘存储、磁盘存储介质或者其他磁存储设备、或者能够用于携带或存储具有指令或数据结构形式的期望的程序代码并能够由计算机存取的任何其他介质。此外。任何连接可以适当的成为计算机可读介质。例如,如果软件是使用同轴电缆、光纤光缆、双绞线、数字用户线(DSL)或者诸如红外线、无线电和微波之类的无线技术从网站、服务器或者其他远程源传输的,那么同轴电缆、光纤光缆、双绞线、DSL或者诸如红外线、无线和微波之类的无线技术包括在所属介质的定影中。如本申请所使用的,盘(Disk)和碟(disc)包括压缩光碟(CD)、激光碟、光碟、数字通用光碟(DVD)、软盘和蓝光光碟,其中盘通常磁性的复制数据,而碟则用激光来光学的复制数据。上面的组合也应当包括在计算机可读介质的保护范围之内。
总之,以上所述仅为本申请技术方案的实施例而已,并非用于限定本申请的保护范围。凡在本申请的精神和原则之内,所作的任何修改、等同替换、改进等,均应包含在本申请的保护范围之内。

Claims (32)

  1. 一种自动驾驶安全评估的方法,其特征在于,包括:
    获取r b和r c,其中,r b为第一计量单位内车辆的影子驾驶模式的风险值,r c为所述第一计量单位内所述车辆基于预设路线的自动驾驶模式的风险值,所述影子驾驶模式为所述车辆基于实时路线的自动驾驶模式,所述实时路线为根据处于人工驾驶模式的所述车辆的位置和运动参数预测的路线,所述第一计量单位为时间段或路程;
    根据r b确定R B,R B为所述车辆在多个计量单位内的影子驾驶模式的风险值,所述多个计量单位包括所述第一计量单位;
    根据r c确定R C,R C为所述车辆在所述多个计量单位内的基于预设路线的自动驾驶模式的风险值;
    其中,R B和R C用于确定所述车辆的自动驾驶模式的安全性是否符合要求。
  2. 根据权利要求1所述的方法,其特征在于,所述方法还包括:
    计算R B与R C之和;
    当所述R B与R C之和小于或等于自动驾驶风险阈值时,确定所述车辆的自动驾驶模式的安全性符合要求。
  3. 根据权利要求2所述的方法,其特征在于,所述自动驾驶风险阈值为a(R A+R′ A),其中,a为风险容忍系数,R A为所述车辆在所述多个计量单位内的人工驾驶的模式的风险值,所述车辆在所述多个计量单位内通过的路段为第一路段,R′ A为所述车辆在所述多个计量单位之外基于人工驾驶模式通过所述第一路段的风险值。
  4. 根据权利要求1至3中任一项所述的方法,其特征在于,所述根据r b确定R B,以及所述根据r c确定R C,包括:
    根据
    Figure PCTCN2019072074-appb-100001
    确定R B或R C,其中,event表示风险场景,s event表示所述风险场景中事故的严重程度的系数,X表示所述多个计量单位,当r i表示r b时,R i表示R B,当r i表示r c时,R i表示R C
  5. 根据权利要求1至4中任一项所述的方法,其特征在于,所述获取r b和r c,包括:
    当风险场景为碰撞场景时,根据下列参数中的至少一个确定r b和r c:所述车辆相对于碰撞物的速度Δv、基于所述碰撞物的运动轨迹预估的所述车辆与所述碰撞物发生碰撞所需的时间TTC、基于所述车辆与所述碰撞物之间的距离预估的所述车辆的刹车时间THW、以及所述车辆与所述碰撞物之间的距离,其中,Δv与r b或r c正相关,TTC与r b或r c负相关,THW与r b或r c负相关,所述车辆与所述碰撞物之间的距离与r b或r c负相关。
  6. 根据权利要求5所述的方法,其特征在于,
    所述根据下列参数中的至少一个确定r b和r c:所述车辆相对于碰撞物的速度Δv、基于所述碰撞物的运动轨迹预估的所述车辆与所述碰撞物发生碰撞所需的时间TTC、基于所述车辆与所述碰撞物之间的距离预估的所述车辆的刹车时间THW、以及所述车辆与所述碰撞物之间的距离,包括:
    根据r SG1=f(t) max确定r b和r c,r SG1为碰撞场景中的r b或r c,其中, t∈{t event,h(t event,obj)=1},f(t)=Δv(t)·exp(-min{r 0TTC(t),r 1THW(t),r 2s}),{t event,h(t event,obj)=1}表示发生碰撞风险的时间段,TTC(t)表示t时刻的TTC,THW(t)表示t时刻的THW,
    Figure PCTCN2019072074-appb-100002
    r 0、r 1和r 2分别为归一化参数,v为t时刻所述车辆的速度,Δv(t)为t时刻的Δv,exp为以自然常数为底的指数,s表示所述车辆与所述碰撞物之间的距离。
  7. 根据权利要求1至4中任一项所述的方法,其特征在于,所述获取r b和r c,包括:
    当风险场景为交通信号灯场景时,根据所述车辆的速度以及所述车辆到交通信号灯的距离确定r b和r c,其中,所述车辆的速度与r b或r c正相关,所述车辆到交通信号灯的距离与r b或r c负相关。
  8. 根据权利要求7所述的方法,其特征在于,
    所述根据所述车辆的速度以及所述车辆与交通信号灯的距离确定r b和r c,包括:
    根据r SG2=f(t) max确定r b和r c,r SG2为交通信号灯场景中的r b或r c,其中,t∈{t event,h(t event,Light)=1},
    Figure PCTCN2019072074-appb-100003
    r 3为归一化参数,{t event,h(t event,Light)=1}表示交发生违反交通信号灯风险的时间段,v(t)表示所述车辆在t时刻的速度,g为重力加速度常数,s(t)表示t时刻所述车辆到交通信号灯的距离。
  9. 根据权利要求1至4中任一项所述的方法,其特征在于,所述获取r b和r c,包括:
    当风险场景为冲出道路边界场景时,根据预估的所述车辆冲出车道时间确定r b和r c,其中,所述预估的所述车辆冲出车道时间与r b或r c负相关。
  10. 根据权利要求9所述的方法,其特征在于,所述根据预估的所述车辆冲出车道时间确定r b和r c,包括:
    根据
    Figure PCTCN2019072074-appb-100004
    确定r b和r c,r SG3为冲出道路边界场景中的r b或r c,其中,TLC threshold表示所述车辆冲出道路所需时间的临界值,
    Figure PCTCN2019072074-appb-100005
    e为自然常数,t lc为所述预估的所述车辆冲出车道时间,y为所述车辆在道路边界的横向偏移量,W road为道路的宽度,v long为所述车辆的纵向速度,v lat为所述车辆的横向速度,φ-φ d表示所述车辆的航向与道路切线方向的夹角,r 4为归一化参数。
  11. 根据权利要求1至4中任一项所述的方法,其特征在于,所述获取r b和r c,包括:
    当风险场景为横向加速度行驶场景时,根据下列参数中的至少一个确定r b和r c:所述车辆的纵向加速度、所述车辆的横向加速度以及所述车辆的速度,其中,所述车辆的纵向加速度与r b或r c正相关,所述车辆的横向加速度与r b或r c正相关,所述车辆的速度与r b或r c正相关。
  12. 根据权利要求11所述的方法,其特征在于,所述根据下列参数中的至少一个确定r b和r c:所述车辆的纵向加速度、所述车辆的横向加速度以及所述车辆的速度,包括:
    根据r SG4=r 5确定r b和r c,r SG4为非常规行驶场景中的r b或r c,其中,(|a long|>a threshold 1)||(|a lat|>a threshold 2)||(v≥v safe),a long表示所述车辆的纵向加速度,a lat表示所述车辆的横向加速度,a threshold 1表示安全行驶状态下横向加速度的阈值,a threshold 2表示安全行驶状态下纵向加速度的阈值,v表示所述车辆的速度,v safe表示安全行驶状态下的速度阈值,r 5为归一化参数。
  13. 一种自动驾驶安全评估的装置,其特征在于,包括处理单元和获取单元,
    所述获取单元用于获取r b和r c,其中,r b为第一计量单位内车辆的影子驾驶模式的风险值,r c为所述第一计量单位内所述车辆基于预设路线的自动驾驶模式的风险值,所述影子驾驶模式为所述车辆基于实时路线的自动驾驶模式,所述实时路线为根据处于人工驾驶模式的所述车辆的位置和运动参数预测的路线,所述第一计量单位为时间段或路程;
    所述处理单元用于根据所述获取单元获取的r b确定R B,R B为所述车辆在多个计量单位内的影子驾驶模式的风险值,所述多个计量单位包括所述第一计量单位;
    所述处理单元还用于根据所述获取单元获取的根据r c确定R C,R C为所述车辆在所述多个计量单位内的基于预设路线的自动驾驶模式的风险值;
    其中,R B和R C用于确定所述车辆的自动驾驶模式的安全性是否符合要求。
  14. 根据权利要求13所述的装置,其特征在于,所述处理单元还用于:
    计算R B与R C之和;
    当所述R B与R C之和小于或等于自动驾驶风险阈值时,确定所述车辆的自动驾驶模式的安全性符合要求。
  15. 根据权利要求14所述的装置,其特征在于,所述自动驾驶风险阈值为a(R A+R′ A),其中,a为风险容忍系数,R A为所述车辆在所述多个计量单位内的人工驾驶的模式的风险值,所述车辆在所述多个计量单位内通过的路段为第一路段,R′ A为所述车辆在所述多个计量单位之外基于人工驾驶模式通过所述第一路段的风险值。
  16. 根据权利要求13至15中任一项所述的装置,其特征在于,所述处理单元具体用于:
    根据
    Figure PCTCN2019072074-appb-100006
    确定R B或R C,其中,event表示风险场景,s event表示所述风险场景中事故的严重程度的系数,X表示所述多个计量单位,当r i表示r b时,R i表示R B,当r i表示r c时,R i表示R C
  17. 根据权利要求13至16中任一项所述的装置,其特征在于,所述获取单元具体用于:
    当风险场景为碰撞场景时,根据下列参数中的至少一个确定r b和r c:所述车辆相对于碰撞物的速度Δv、基于所述碰撞物的运动轨迹预估的所述车辆与所述碰撞物发生碰撞所需的时间TTC、基于所述车辆与所述碰撞物之间的距离预估的所述车辆的刹车时间THW、以及所述车辆与所述碰撞物之间的距离,其中,Δv与r b或r c正相关,TTC与r b或r c负相关,THW与r b或r c负相关,所述车辆与所述碰撞物之间的距离与r b或r c负相关。
  18. 根据权利要求17所述的装置,其特征在于,所述获取单元具体还用于:
    根据r SG1=f(t) max确定r b和r c,r SG1为碰撞场景中的r b或r c,其中,t∈{t event,h(t event,obj)=1},f(t)=Δv(t)·exp(-min{r 0TTC(t),r 1THW(t),r 2s}),{t event,h(t event,obj)=1}表示发生碰撞风险的时间段,TTC(t)表示t时刻的TTC,THW(t)表示t时刻的THW,
    Figure PCTCN2019072074-appb-100007
    r 0、r 1和r 2分别为归一化参数,v为t时刻所述车辆的速度,Δv(t)为t时刻的Δv,exp为以自然常数为底的指数,s表示所述车辆与所述碰撞物之间的距离。
  19. 根据权利要求13至16中任一项所述的装置,其特征在于,所述获取单元具体用于:
    当风险场景为交通信号灯场景时,根据所述车辆的速度以及所述车辆到交通信号灯的距离确定r b和r c,其中,所述车辆的速度与r b或r c正相关,所述车辆到交通信号灯的距离与r b或r c负相关。
  20. 根据权利要求19所述的装置,其特征在于,所述获取单元具体还用于:
    根据r SG2=f(t) max确定r b和r c,r SG2为交通信号灯场景中的r b或r c,其中,t∈{t event,h(t event,Light)=1},
    Figure PCTCN2019072074-appb-100008
    r 3为归一化参数,{t event,h(t event,Light)=1}表示交发生违反交通信号灯风险的时间段,v(t)表示所述车辆在t时刻的速度,g为重力加速度常数,s(t)表示t时刻所述车辆到交通信号灯的距离。
  21. 根据权利要求13至16中任一项所述的装置,其特征在于,所述获取单元具体用于:
    当风险场景为冲出道路边界场景时,根据预估的所述车辆冲出车道时间确定r b和r c,其中,所述预估的所述车辆冲出车道时间与r b或r c负相关。
  22. 根据权利要求21所述的装置,其特征在于,所述获取单元具体还用于:
    根据
    Figure PCTCN2019072074-appb-100009
    确定r b和r c,r SG3为冲出道路边界场景中的r b或r c,其中,TLC threshold表示所述车辆冲出道路所需时间的临界值,
    Figure PCTCN2019072074-appb-100010
    e为自然常数,t lc为所述预估的所述车辆冲出车道时间,y为所述车辆在道路边界的横向偏移量,W road为道路的宽度,v long为所述车辆的纵向速度,v lat为所述车辆的横向速度,φ-φ d表示所述车辆的航向与道路切线方向的夹角,r 4为归一化参数。
  23. 根据权利要求13至16中任一项所述的装置,其特征在于,所述获取单元具体用于:
    当风险场景为横向加速度行驶场景时,根据下列参数中的至少一个确定r b和r c:所述车辆的纵向加速度、所述车辆的横向加速度以及所述车辆的速度,其中,所述车辆的纵向 加速度与r b或r c正相关,所述车辆的横向加速度与r b或r c正相关,所述车辆的速度与r b或r c正相关。
  24. 根据权利要求23所述的装置,其特征在于,所述获取单元具体还用于:
    根据r SG4=r 5确定r b和r c,r SG4为非常规行驶场景中的r b或r c,其中,(|a long|>a threshold 1)||(|a lat|>a threshold 2)||(v≥v safe),a long表示所述车辆的纵向加速度,a lat表示所述车辆的横向加速度,a threshold 1表示安全行驶状态下横向加速度的阈值,a threshold 2表示安全行驶状态下纵向加速度的阈值,v表示所述车辆的速度,v safe表示安全行驶状态下的速度阈值,r 5为归一化参数。
  25. 一种确定车辆驾驶风险的方法,其特征在于,所述方法包括:
    确定一个时间段内车辆的驾驶模式和风险场景,所述时间段的起始点为第一时刻,所述时间段的终点为第二时刻,所述驾驶模式包括自动驾驶模式和/或人工驾驶模式;
    在当前时刻根据风险场景对应的风险识别方式确定所述时间段内所述驾驶模式的风险值,其中,所述当前时刻位于所述第一时刻之后,且所述当前时刻位于所述第二时刻之前。
  26. 根据权利要求25所述的方法,其特征在于,
    所述在当前时刻根据风险场景对应的风险识别方式确定所述时间段内所述驾驶模式的风险值之前,所述方法还包括:
    确定所述车辆在所述第一时刻的位置和运动参数,其中,车辆在所述第一时刻基于人工驾驶模式行驶;
    根据所述车辆在第一时刻的位置和运动参数预测车辆在第三时刻的位置,所述第三时刻晚于所述第一时刻;
    根据所述第一时刻的位置和所述第三时刻的位置预测所述车辆从所述第一时刻的位置到所述第三时刻的位置的影子驾驶模式,所述影子驾驶模式为基于所述车辆处于人工驾驶模式下的位置和运动参数预测的所述车辆的自动驾驶模式;
    所述在当前时刻根据所述风险场景对应的风险识别方式确定所述时间段内所述驾驶模式的风险值,包括:
    在所述当前时刻根据所述风险场景对应的风险识别方式确定r a、r b和r c,其中,r a为所述时间段内所述人工驾驶模式的风险值,r b为所述时间段内所述影子驾驶模式的风险值,r c为所述时间段内基于预设路线的自动驾驶模式的风险值。
  27. 一种确定车辆驾驶风险的装置,其特征在于,所述装置包括处理单元和存储单元,所述存储单元存储了计算机程序代码,当所述计算机程序代码被所述处理单元或执行时,所述处理单元用于:
    确定一个时间段内车辆的驾驶模式和风险场景,所述时间段的起始点为第一时刻,所述时间段的终点为第二时刻,所述驾驶模式包括自动驾驶模式和/或人工驾驶模式;
    在当前时刻根据风险场景对应的风险识别方式确定所述时间段内所述驾驶模式的风险值,其中,所述当前时刻位于所述第一时刻之后,且所述当前时刻位于所述第二时刻之前。
  28. 根据权利要求27所述的装置,其特征在于,
    所述在当前时刻根据风险场景对应的风险识别方式确定所述时间段内所述驾驶模式的风险值之前,所述处理单元还用于:
    确定所述车辆在所述第一时刻的位置和运动参数,其中,车辆在所述第一时刻基于人工驾驶模式行驶;
    根据所述车辆在第一时刻的位置和运动参数预测车辆在第三时刻的位置,所述第三时刻晚于所述第一时刻;
    根据所述第一时刻的位置和所述第三时刻的位置预测所述车辆从所述第一时刻的位置到所述第三时刻的位置的影子驾驶模式,所述影子驾驶模式为基于所述车辆处于人工驾驶模式下的位置和运动参数预测的所述车辆的自动驾驶模式;
    所述处理单元具体用于:
    在所述当前时刻根据所述风险场景对应的风险识别方式确定r a、r b和r c,其中,r a为所述时间段内所述人工驾驶模式的风险值,r b为所述时间段内所述影子驾驶模式的风险值,r c为所述时间段内基于预设路线的自动驾驶模式的风险值。
  29. 一种自动驾驶安全评估系统,其特征在于,包括风险控制模块、风险策略模块和安全评估模块,
    所述风险控制模块用于执行权利要求1至12中任一项所述的方法,确定自动驾驶模式的风险;
    所述风险策略模块用于根据所述风险控制模块确定的所述自动驾驶模式的风险确定风险控制策略,所述风险控制策略用于减小或消除所述自动驾驶模式的风险;
    所述安全评估策略用于根据所述风险控制模块确定的所述自动驾驶模式的风险评估所述自动驾驶模式的安全性。
  30. 一种自动驾驶安全评估系统,其特征在于,包括风险控制模块、风险策略模块和安全评估模块,
    所述风险控制模块用于执行权利要求25或26所述的方法,确定自动驾驶模式的风险;
    所述风险策略模块用于根据所述风险控制模块确定的所述自动驾驶模式的风险确定风险控制策略,所述风险控制策略用于减小或消除所述自动驾驶模式的风险;
    所述安全评估策略用于根据所述风险控制模块确定的所述自动驾驶模式的风险评估所述自动驾驶模式的安全性。
  31. 一种可读存储介质,包括程序或指令,当所述程序或指令在计算机上运行时,如权利要求1至12中任意一项所述的方法被执行。
  32. 一种可读存储介质,包括程序或指令,当所述程序或指令在计算机上运行时,如权利要求25或26所述的方法被执行。
PCT/CN2019/072074 2018-03-16 2019-01-17 自动驾驶安全评估方法、装置和系统 WO2019174397A1 (zh)

Priority Applications (5)

Application Number Priority Date Filing Date Title
EP19768393.1A EP3745378B1 (en) 2018-03-16 2019-01-17 Automatic driving safety assessment method, device and system
KR1020207026721A KR102574105B1 (ko) 2018-03-16 2019-01-17 자율 주행 안전성 평가 방법, 디바이스, 및 시스템
JP2020549588A JP7227267B2 (ja) 2018-03-16 2019-01-17 自動運転安全性評価方法、装置、およびシステム
BR112020017519-6A BR112020017519A2 (pt) 2018-03-16 2019-01-17 Método, e dispositivo de avaliação de segurança de direção autônoma e de risco de direção, sistema de avaliação de segurança de direção autônoma, e meio legível por computador
US17/018,505 US11872999B2 (en) 2018-03-16 2020-09-11 Self-driving safety evaluation method, apparatus, and system

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201810220891.XA CN110276985B (zh) 2018-03-16 2018-03-16 自动驾驶安全评估方法、装置和系统
CN201810220891.X 2018-03-16

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US17/018,505 Continuation US11872999B2 (en) 2018-03-16 2020-09-11 Self-driving safety evaluation method, apparatus, and system

Publications (1)

Publication Number Publication Date
WO2019174397A1 true WO2019174397A1 (zh) 2019-09-19

Family

ID=67908627

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/072074 WO2019174397A1 (zh) 2018-03-16 2019-01-17 自动驾驶安全评估方法、装置和系统

Country Status (7)

Country Link
US (1) US11872999B2 (zh)
EP (1) EP3745378B1 (zh)
JP (1) JP7227267B2 (zh)
KR (1) KR102574105B1 (zh)
CN (1) CN110276985B (zh)
BR (1) BR112020017519A2 (zh)
WO (1) WO2019174397A1 (zh)

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110781578A (zh) * 2019-09-23 2020-02-11 同济大学 一种基于事故场景的智能网联车算法的测试及评价方法
CN111783226A (zh) * 2020-06-29 2020-10-16 北京百度网讯科技有限公司 生成自动驾驶场景度量参数的方法、装置、电子设备和存储介质
CN112816226A (zh) * 2021-01-06 2021-05-18 北京智能车联产业创新中心有限公司 基于可控交通流的自动驾驶测试系统及方法
CN113200054A (zh) * 2021-06-21 2021-08-03 清华大学 一种自动驾驶接管的路径规划方法及系统
CN113642108A (zh) * 2021-08-11 2021-11-12 北京航空航天大学 一种基于重要度抽样的无人车十字路口通行场景关键测试案例生成方法
FR3114788A1 (fr) * 2020-10-07 2022-04-08 Psa Automobiles Sa Procédé et dispositif de détermination d’un indicateur de comportement latéral dangereux d’un véhicule autonome circulant sur une voie de circulation
CN115150194A (zh) * 2022-08-12 2022-10-04 国汽智控(北京)科技有限公司 防御策略的确定方法、装置、设备及存储介质
CN115465288A (zh) * 2022-08-12 2022-12-13 重庆长安汽车股份有限公司 自动驾驶车辆的控制方法、装置、车辆及存储介质
CN116052471A (zh) * 2023-01-09 2023-05-02 内蒙古大学 一种车辆过弯冲撞路侧护栏的风险评估方法及系统
US20230347882A1 (en) * 2022-04-29 2023-11-02 Gm Cruise Holdings Llc Baselining autonomous vehicle safety using driving data of human autonomous vehicle operators
CN118227510A (zh) * 2024-05-22 2024-06-21 中国汽车技术研究中心有限公司 自动导航辅助驾驶系统评价方法、设备和存储介质

Families Citing this family (35)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2019191893A (ja) * 2018-04-24 2019-10-31 本田技研工業株式会社 車両制御装置及びプログラム
US20220385559A1 (en) * 2019-07-03 2022-12-01 Nippon Telegraph And Telephone Corporation Topology estimation system, traffic generation apparatus, and traffic generation method
CN112556700A (zh) * 2019-09-25 2021-03-26 华为技术有限公司 地图级别指示方法、地图级别获取方法及相关产品
CN110751825B (zh) * 2019-10-29 2021-09-14 北京百度网讯科技有限公司 编队行驶的避让方法、装置、设备和计算机可读存储介质
CN110782706B (zh) * 2019-11-06 2022-05-06 腾讯科技(深圳)有限公司 智能车辆驾驶风险的预警方法和装置
CN110853389B (zh) * 2019-11-21 2022-03-18 白犀牛智达(北京)科技有限公司 一种适用于无人物流车的路测监控系统
CN112834234A (zh) * 2019-11-25 2021-05-25 武汉理工大学 一种判定目标假车碰撞的方法及系统
US11385656B2 (en) 2020-01-22 2022-07-12 Huawei Technologies Co., Ltd. System, device and method of identifying and updating the operational design domain of an autonomous vehicle
CN111710187A (zh) * 2020-05-19 2020-09-25 腾讯科技(深圳)有限公司 确定车辆驾驶风险的方法、装置、介质及电子设备
CN111611709B (zh) * 2020-05-20 2023-09-15 阿波罗智能技术(北京)有限公司 用于控制仿真车辆的方法、装置、电子设备和存储介质
CN111717221B (zh) * 2020-05-29 2022-11-11 重庆大学 自动驾驶接管风险评估和人机友好预警方法及预警系统
CN113752827B (zh) * 2020-06-05 2024-03-01 北京京东乾石科技有限公司 一种无人车速度控制方法、装置、无人车及存储介质
CN111814904B (zh) * 2020-07-22 2021-05-28 同济大学 一种自动驾驶道路测试驾驶模式辨别方法及系统
CN112100030B (zh) * 2020-09-15 2023-07-28 北京百度网讯科技有限公司 自动驾驶技术评估的方法、装置、计算机系统和存储介质
CN112232581B (zh) * 2020-10-26 2024-04-12 腾讯科技(深圳)有限公司 驾驶风险的预测方法、装置、电子设备及存储介质
CN112525550B (zh) * 2020-12-03 2024-07-19 苏州测迅智能汽车科技有限公司 一种小型化智能车辆感知系统在环测试平台
CN112729862B (zh) * 2020-12-11 2023-07-25 国汽(北京)智能网联汽车研究院有限公司 无人驾驶的驾驶行为安全评估方法、装置及电子设备
CN114661574A (zh) * 2020-12-23 2022-06-24 北京百度网讯科技有限公司 样本偏差数据的获取方法、装置和电子设备
CN113158560B (zh) * 2021-04-09 2024-02-09 中国科学院合肥物质科学研究院 一种基于场景对抗的智能驾驶车辆自主能力测试方法
CN116761998A (zh) * 2021-05-28 2023-09-15 深圳市大疆创新科技有限公司 一种车辆系统的安全测试方法和测试用车辆
CN113419233A (zh) * 2021-06-18 2021-09-21 阿波罗智能技术(北京)有限公司 感知效果的测试方法、装置和设备
CN113658451B (zh) * 2021-07-30 2022-11-22 三一专用汽车有限责任公司 车辆碰撞预警的控制方法、装置、车辆和可读存储介质
CN113479201B (zh) * 2021-08-20 2022-08-12 燕山大学 考虑驾驶人反应能力的换道场景车辆风险动态评价方法
CN114627640B (zh) * 2022-01-21 2023-07-28 中国科学院深圳先进技术研究院 一种智能网联汽车行驶策略的动态演化方法
KR102434475B1 (ko) * 2022-01-21 2022-08-19 주식회사 모라이 자율 주행 알고리즘의 성능 평가 방법 및 시스템
CN114613131B (zh) * 2022-03-01 2023-03-28 北京航空航天大学 一种基于安全裕度的个性化前向碰撞预警方法
CN114724366B (zh) * 2022-03-29 2023-06-20 北京万集科技股份有限公司 辅助驾驶方法、装置、设备和存储介质
CN114454876B (zh) * 2022-04-13 2022-07-26 所托(杭州)汽车智能设备有限公司 车辆碰撞确定方法、检测装置及存储介质
CN114973669B (zh) * 2022-05-23 2023-09-26 江苏智行未来汽车研究院有限公司 基于车路协同的公路危险环境预警方法、装置及介质
CN114822042B (zh) * 2022-06-28 2022-09-02 深圳市华耀商品检验有限公司 一种用于车载终端检测的信息安全测试管理系统及方法
CN115146539B (zh) * 2022-07-12 2023-09-22 小米汽车科技有限公司 车辆的安全评价参数确定方法、装置和车辆
CN115447616B (zh) * 2022-10-26 2024-05-17 重庆长安汽车股份有限公司 车辆驾驶客观指标的生成方法及装置
CN117093811B (zh) * 2023-07-17 2024-05-10 上海智能汽车融合创新中心有限公司 一种自动驾驶安全能力评价方法、装置及设备
CN117315938B (zh) * 2023-10-19 2024-08-20 广州市城市规划勘测设计研究院有限公司 一种高速公路车辆换道风险估算方法、介质及设备
CN117601903B (zh) * 2024-01-19 2024-04-26 福思(杭州)智能科技有限公司 车辆行驶场景的生成方法、装置和存储介质及电子设备

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2009274594A (ja) * 2008-05-15 2009-11-26 Hitachi Ltd 車線変更支援装置
JP2011134103A (ja) * 2009-12-24 2011-07-07 Aisin Aw Co Ltd 案内装置、案内方法、及び案内プログラム
JP2012160128A (ja) * 2011-02-02 2012-08-23 Toyota Motor Corp 進路生成装置
CN103917432A (zh) * 2011-11-08 2014-07-09 丰田自动车株式会社 车辆的行驶轨迹控制装置
CN106164998A (zh) * 2014-04-10 2016-11-23 三菱电机株式会社 路径预测装置
CN106796759A (zh) * 2014-12-26 2017-05-31 日立汽车系统株式会社 车辆控制系统

Family Cites Families (24)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2304511B1 (en) 2008-06-20 2013-05-22 Toyota Jidosha Kabushiki Kaisha Driving assistance apparatus
DE102009052590A1 (de) * 2009-11-10 2011-05-12 Valeo Schalter Und Sensoren Gmbh Vorrichtung und Verfahren zur Überwachung und Signalisierung von Verkehrssituationen und Betriebszuständen in einem Fahrzeug und im Umfeld des Fahrzeugs
US20110190972A1 (en) * 2010-02-02 2011-08-04 Gm Global Technology Operations, Inc. Grid unlock
US9173100B2 (en) * 2011-11-16 2015-10-27 Autoconnect Holdings Llc On board vehicle network security
DE102012215064A1 (de) * 2012-08-24 2014-02-27 Robert Bosch Gmbh Verfahren und Vorrichtung zum Ermitteln einer Gefahrenquelle einer Fahrstrecke
US9633564B2 (en) 2012-09-27 2017-04-25 Google Inc. Determining changes in a driving environment based on vehicle behavior
JP2014193691A (ja) * 2013-03-29 2014-10-09 Hitachi Automotive Systems Ltd 車両の運動制御装置
DE102013211607A1 (de) * 2013-06-20 2014-12-24 Robert Bosch Gmbh Verfahren und Vorrichtung zum Betreiben eines Fahrzeugs
CN104670222B (zh) 2013-11-29 2017-07-25 财团法人车辆研究测试中心 滑动模式的行驶路径投票策略模块及驾驶控制系统与方法
US9650051B2 (en) 2013-12-22 2017-05-16 Lytx, Inc. Autonomous driving comparison and evaluation
JP6204865B2 (ja) * 2014-03-31 2017-09-27 日立オートモティブシステムズ株式会社 車両の運動制御システム、車両、および、プログラム
CN104002808A (zh) * 2014-06-05 2014-08-27 大连理工大学 一种汽车主动防碰撞自动制动控制系统及工作方法
JP6431594B2 (ja) 2015-03-31 2018-11-28 日立オートモティブシステムズ株式会社 自動運転制御装置
JP6803657B2 (ja) 2015-08-31 2020-12-23 日立オートモティブシステムズ株式会社 車両制御装置および車両制御システム
DE102015220360A1 (de) * 2015-10-20 2017-04-20 Robert Bosch Gmbh Verfahren zur Auswahl einer optimierten Trajektorie
KR20170053799A (ko) * 2015-11-06 2017-05-17 고려대학교 산학협력단 자율 주행 차량의 안전성 제공 장치 및 방법
CN105573323A (zh) * 2016-01-12 2016-05-11 福州华鹰重工机械有限公司 自动驾驶轨迹生成方法及装置
EP3239686B1 (en) * 2016-04-26 2024-09-18 Walter Steven Rosenbaum Method for determining driving characteristics of a vehicle
CN105956268B (zh) * 2016-04-29 2018-01-02 百度在线网络技术(北京)有限公司 应用于无人驾驶汽车的测试场景构建方法和装置
US10019009B2 (en) * 2016-06-15 2018-07-10 Allstate Insurance Company Vehicle control systems
US10007854B2 (en) * 2016-07-07 2018-06-26 Ants Technology (Hk) Limited Computer vision based driver assistance devices, systems, methods and associated computer executable code
CN106218639B (zh) * 2016-07-20 2019-01-11 百度在线网络技术(北京)有限公司 无人驾驶车辆、用于控制无人驾驶车辆的方法和装置
CN106347359B (zh) 2016-09-14 2019-03-12 北京百度网讯科技有限公司 用于操作自动驾驶车辆的方法和装置
WO2018053175A1 (en) * 2016-09-14 2018-03-22 Nauto Global Limited Systems and methods for near-crash determination

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2009274594A (ja) * 2008-05-15 2009-11-26 Hitachi Ltd 車線変更支援装置
JP2011134103A (ja) * 2009-12-24 2011-07-07 Aisin Aw Co Ltd 案内装置、案内方法、及び案内プログラム
JP2012160128A (ja) * 2011-02-02 2012-08-23 Toyota Motor Corp 進路生成装置
CN103917432A (zh) * 2011-11-08 2014-07-09 丰田自动车株式会社 车辆的行驶轨迹控制装置
CN106164998A (zh) * 2014-04-10 2016-11-23 三菱电机株式会社 路径预测装置
CN106796759A (zh) * 2014-12-26 2017-05-31 日立汽车系统株式会社 车辆控制系统

Cited By (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110781578B (zh) * 2019-09-23 2024-03-29 同济大学 一种基于事故场景的智能网联车算法的测试及评价方法
CN110781578A (zh) * 2019-09-23 2020-02-11 同济大学 一种基于事故场景的智能网联车算法的测试及评价方法
CN111783226A (zh) * 2020-06-29 2020-10-16 北京百度网讯科技有限公司 生成自动驾驶场景度量参数的方法、装置、电子设备和存储介质
CN111783226B (zh) * 2020-06-29 2024-03-19 北京百度网讯科技有限公司 生成自动驾驶场景度量参数的方法、装置、电子设备和存储介质
FR3114788A1 (fr) * 2020-10-07 2022-04-08 Psa Automobiles Sa Procédé et dispositif de détermination d’un indicateur de comportement latéral dangereux d’un véhicule autonome circulant sur une voie de circulation
CN112816226B (zh) * 2021-01-06 2023-03-14 北京智能车联产业创新中心有限公司 基于可控交通流的自动驾驶测试系统及方法
CN112816226A (zh) * 2021-01-06 2021-05-18 北京智能车联产业创新中心有限公司 基于可控交通流的自动驾驶测试系统及方法
CN113200054A (zh) * 2021-06-21 2021-08-03 清华大学 一种自动驾驶接管的路径规划方法及系统
CN113642108A (zh) * 2021-08-11 2021-11-12 北京航空航天大学 一种基于重要度抽样的无人车十字路口通行场景关键测试案例生成方法
CN113642108B (zh) * 2021-08-11 2023-11-10 北京航空航天大学 一种无人车十字路口通行场景关键测试案例生成方法
US20230347882A1 (en) * 2022-04-29 2023-11-02 Gm Cruise Holdings Llc Baselining autonomous vehicle safety using driving data of human autonomous vehicle operators
CN115150194A (zh) * 2022-08-12 2022-10-04 国汽智控(北京)科技有限公司 防御策略的确定方法、装置、设备及存储介质
CN115465288A (zh) * 2022-08-12 2022-12-13 重庆长安汽车股份有限公司 自动驾驶车辆的控制方法、装置、车辆及存储介质
CN116052471A (zh) * 2023-01-09 2023-05-02 内蒙古大学 一种车辆过弯冲撞路侧护栏的风险评估方法及系统
CN116052471B (zh) * 2023-01-09 2024-03-19 内蒙古大学 一种车辆过弯冲撞路侧护栏的风险评估方法及系统
CN118227510A (zh) * 2024-05-22 2024-06-21 中国汽车技术研究中心有限公司 自动导航辅助驾驶系统评价方法、设备和存储介质

Also Published As

Publication number Publication date
JP2021516816A (ja) 2021-07-08
CN110276985A (zh) 2019-09-24
KR102574105B1 (ko) 2023-09-06
JP7227267B2 (ja) 2023-02-21
EP3745378A1 (en) 2020-12-02
KR20200118882A (ko) 2020-10-16
EP3745378A4 (en) 2021-03-24
US11872999B2 (en) 2024-01-16
CN110276985B (zh) 2020-12-15
EP3745378B1 (en) 2022-06-01
US20200406911A1 (en) 2020-12-31
BR112020017519A2 (pt) 2020-12-22

Similar Documents

Publication Publication Date Title
WO2019174397A1 (zh) 自动驾驶安全评估方法、装置和系统
US11430071B2 (en) Navigation based on liability constraints
JP6708793B2 (ja) 課せられる責任の制約を伴うナビゲーションシステム
US20220227367A1 (en) Systems and methods for vehicle navigation
US11077845B2 (en) Systems and methods for navigating a vehicle
US20210094577A1 (en) Systems and Methods for Navigating with Safe Distances
CN110402371B (zh) 用于利用感测不确定性导航的系统和方法
CN110675656B (zh) 一种基于瞬时风险识别的智能车辆换道预警方法
US20150360697A1 (en) System and method for managing dangerous driving index for vehicle
US20230286536A1 (en) Systems and methods for evaluating domain-specific navigation system capabilities

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: 19768393

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2019768393

Country of ref document: EP

Effective date: 20200827

ENP Entry into the national phase

Ref document number: 2020549588

Country of ref document: JP

Kind code of ref document: A

ENP Entry into the national phase

Ref document number: 20207026721

Country of ref document: KR

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

REG Reference to national code

Ref country code: BR

Ref legal event code: B01A

Ref document number: 112020017519

Country of ref document: BR

ENP Entry into the national phase

Ref document number: 112020017519

Country of ref document: BR

Kind code of ref document: A2

Effective date: 20200827