Research On Human-machine Collaborative Decision-making And Planning Technology For Intelligent Vehicles On Road Environment

Intelligent driving system can improve driving safety and reduce traffic accidents effectively.However,the current intelligent driving system still has many limitations,and faces many challenges in environmental understanding,decision,planning and control.Drivers and intelligent driving systems can be complementary.Drivers have rich experience and common sense in environmental understanding.Intelligent driving systems have the ability to continuously monitor the surrounding environment and can assist to drive safely to reduce the driver’s workload.How to integrate the advantages of drivers and intelligent driving system through human-machine cooperative driving technology has become an important research topic of intelligent driving technology.Based on the problem of human-machine cooperative driving of vehicles,this paper focuses on the human-machine intelligent interaction at the decision level and planning level.Based on the comprehensive modeling and analysis of "human-vehicleroad" information,this paper carries out the research on analysis of the vehicle safety motion safety boundary,the human-machine cooperative decision-making algorithm based on the safety boundary and human-machine cooperative trajectory planning method based on driving intention understanding.The main work and innovations are as follows.(1)Aiming at solving the problem of risk assessment in the process of humanmachine cooperative driving,a method of calculating vehicle safety motion boundaries is proposed.Firstly,the driving scene is modeled.Secondly,the vehicle movement is modeled.Thirdly,based on the understanding of environmental semantics,the safe distance modeling of different driving behaviors is proposed.At last,the safety motion boundaries of different driving behaviors are generated using the comprehensive information of "human-vehicle-road",including the vehicle state,the vehicle motion constraint,the driver control constraint,the road constraint and the collision avoidance constraint.The safety boundaries are important bases for judging the safety of driving behavior and determining the timing of warning intervention.The algorithm has been fully verified in the Prescan simulation environment.(2)Aiming at solving the problem of human-machine collaboration at the behavior decision-making level,based on the above research on vehicle movement safety boundary,a hazard warning and a decision aid algorithm is proposed based on rules and the "person-vehicle-road" comprehensive information.The hazard warning algorithm proposed in this paper can remind the driver to correct the operation,so that the vehicle can enter the safe driving state again.The decision aid algorithm proposed in this paper can evaluate the safety and optimality of candidate lane change behaviors and obstacle avoidance behaviors.The decision aid algorithm can also make decisions on speed adjustment strategies and timing of lane change.The Prescan simulation software and the real vehicle test platform are used to fully verify the effectiveness of the above two methods.(3)Aiming at solving the problem of man-machine cooperation at the trajectory planning layer,a human-machine cooperative driving trajectory planning algorithm combining driving intention and "man-vehicle-road" comprehensive information is proposed.While respecting the driver’s intent,the machine can perform trajectory planning in a human-machine collaborative manner,and realize man-machine shared steering control based on torque superposition.The actual vehicle test shows that the algorithm has the ability to assist the driver in steering and braking in case of emergency,and has the ability to realize lane change assistance according to the driver’s lane change intention.