Evaluation Of Cooperative Driving Behavior Based On The Virtual Driving

As the development of autopilot technology,the way to drive vehicles of has been also changed.Autopilot vehicles will be always in the phase of cooperative driving for a long period of time before they have achieved full drone driving.The new driving problems will be caused by the new way of driving.The traditional driving behavior theory can not describe the new cooperative driving behaviors.Therefore,based on the study of the traditional virtual driving experiment platform and the existing research on the human-motor co-drive behavior,this paper sets up a man-machine mutual driving virtual experiments platform.The platform designs an early warning experiment of autopilot system,a take-over experiment when the platform is failed and an adaptive experiment when the driver taking over the platform.Three groups of experiments are organized respectively.The co-driven behavior of the switched man-machine was studied and evaluated from three perspectives.It mainly includes the following aspects:(1)The cooperative driving virtual experiment platform was built by the low-cost hardware and Matlab/3D Animation Toolbox.The main contents of the platform include: Firstly,establishing a more realistic virtual driving scene and introducing the vehicle's five-degree-of-freedom dynamics model into driving simulation.Then,finishing the two driving modes of the master vehicle of the experimental platform and realizing the switching between the two driving modes.Finally,a series of human-machine interaction modules were designed to lay the human-machine co-drive experimental basis.(2)We take a research on a scene of switching the cooperative driving which is the system was taken over after the early warning.We design the warning experiments and organize the experimental personnel to complete related experiments.Based on the experimental data,the driver's operation behavior after taking over and the driving status of the vehicle are analyzed,and the safety hazards taken over under this scene are evaluated.At the same time,the reaction time of the driver from the time of early warning to the time of taking over was counted,and the probability distribution curve of the takeover response time under two kinds of early warning conditions was fitted.The results provide some reference for the design of man-machine co-drive vehicles in the future.(3)We also take a research on switching the other kind of take-over scene—taken after failure of the autopilot system.Firstly,we design an automatic driving failure generator to simulate the failure scene of the autopilot longitudinal control system and the failure-taken-over experiment.Then,by analyzing the experimental data,the indicators that define the driver's failure to judge the automatic control longitudinal control system are discovered.Finally,the conditions for the driver to successfully take over after the failure of the autopilot system are given.(4)After studying the two common take-over scenes of the switched man-machine co-drive,the lateral control ability after the driver took over was studied.Two groups of experiments are taken to compare the behavior of the drivers when they changing the lane at the speed of 30 km/h and 60 km/h.The experimental data are analyzed from two aspects of the driver's operation and the vehicle's track.The results show that the driver's lateral control ability after taking over is lower than the lateral control ability during manual driving,but this phenomenon will be weaken or disappeared through repeated operations.