基于驾驶模拟器的人机共驾策略评价

Strategy Evaluation for Human-machine Shared Driving based on Driving Simulator

人机共驾是一种驾驶员与智能系统通过合作,共享车辆控制权并实现协同驾驶的驾驶方式。该方式相比手动驾驶模式,能够降低事故率,并在一定程度上解决自动驾驶带来的伦理问题。目前面向人机共驾的研究仍处于初步阶段,开展实车实验具有较大的安全隐患,且现有人机共驾策略的评价方法有所欠缺。针对以上问题,本文基于自主设计的人机共驾驾驶模拟器进行实验,对人机共驾策略评价方法进行研究。通过客观指标的主成分分析以及主观问卷调查表的相关性分析,提出人机共驾策略主客观评价指标,并进行理论论证,证明了指标与评价方法的有效性。

Human-machine shared driving is a kind of driving mode in which the driver and the intelligent system share the control au⁃thority of the vehicle through cooperation and realize cooperative driving.To some extent,compared with manual driving mode,this method can reduce the accident rate and solve the ethical problems caused by automatic driving.At present,the research on human-machine shared driving is still in the preliminary stage,carrying out real car experiments has great safety risks,and the evaluation methods for human-machine shared driving strategies are lacking.Aiming at the above problems,this paper conducts experiments based on the self-designed human-machine shared driving simulator,and studies the evaluation method of human-machine shared driving strategy.Through principal component analysis(PCA)of objective indicators and correlation analysis with subjective questionnaire,the subjective and objec⁃tive evaluation indicators of human-machine shared driving strategy are proposed,and theoretical demonstration is carried out to prove the effectiveness of the indicators and evaluation methods.