Speed Optimization For Connected Autonomous Vehicles Under High-speed Vehicle-pedestrian Conflicts

As the vulnerable road traffic user,pedestrian safety has become a hot issue in traffic safety.At present,the researches on vehicle active safety control to alleviate vehicle-pedestrian conflict mainly focus on urban low-speed traffic scenes.In the scene of rapid traffic,the aggravation of vehicle-pedestrian conflict will multiply the threat to pedestrian.Vehiclepedestrian collision avoidance in this scene requires more adequate safety margins,which will put forward higher requirements for pedestrian trajectory prediction and vehicle speed control.Therefore,it is urgent to explore the pedestrian active anti-collision control strategy of highspeed vehicles with the aid of V2 X communication technology.According to the characteristics of vehicle-pedestrian conflict in rapid traffic,this paper firstly constructs vehicle-pedestrian conflict data sets on expressway and freeway,then proposes a vehicle-pedestrian conflict time window prediction model combining pedestrian crossing time prediction and real-time monitoring of motion trajectory,and finally constructs a speed optimization model for connected autonomous vehicles based on vehicle-pedestrian conflict time window framework,so as to guide vehicles to drive safely and smoothly to the pedestrian crossing point and ensure safety of vehicle and pedestrian.Firstly,through the analysis of characteristics of vehicle-pedestrian conflict in rapid traffic scenes,the operation mechanism of connected autonomous vehicle-pedestrian conflict is determined.Secondly,the vehicle-pedestrian conflict data sets of expressway and freeway are constructed by means of field collection and simulation output,which are used for the calibration,verification and testing of subsequent models.Next,by introducing the relaxation time parameter,a vehicle-pedestrian conflict time window prediction model combining pedestrian crossing time prediction and real-time monitoring of motion trajectory is proposed,and its performance is verified and tested.Finally,taking safety as the basic premise and taking into account the objectives of traffic efficiency and fuel economy,a speed optimization model for connected autonomous vehicles under conflict time window framework is constructed,and the model is tested from three indicators: safety,traffic efficiency and fuel economy by defining three control strategies.The results show that the vehicle-pedestrian conflict time window prediction model can effectively reduce the influence of pedestrian crossing randomness on trajectory prediction,provide sufficient safety margin while taking into account the reset times of prediction results,and greatly improve pedestrian safety with a small loss of traffic efficiency.The speed optimization model for connected autonomous vehicles based on the conflict time window framework can significantly improve the safety of vehicles and pedestrians with a slight sacrifice of traffic efficiency and fuel economy.Thus,with the help of the framework of vehicle-pedestrian conflict time window,methods of pedestrian trajectory prediction and speed optimization for connected autonomous vehicles are studied to solve the vehicle-pedestrian conflict problems in rapid transit scenarios,which is beneficial to improve the safety of pedestrians and vehicles in rapid transit,and give consideration to traffic efficiency and fuel economy.It provides theoretical basis and technical support for the safe,efficient and environmental-friendly operation of rapid traffic.