智能汽车人机共驾风险认知可信度建模及冲突消解策略研究

驾驶人和智能系统共同操纵车辆的局面在智能汽车技术发展过程中将长期存在，现有技术仍不能保证智能汽车行驶的绝对安全，人机共驾过程的操控舒适性仍待提高，严重制约该技术的发展与应用。为此，本项目拟以提高智能汽车的行驶安全性和操控舒适性为总体目标，围绕“考虑驾驶人主观风险感受的二维行车风险建模”和“主观与客观行车风险评估层面的人机认知冲突”两大关键科学问题，深入研究行车风险源及其梯度变化，提出基于驾驶人认知特性的人-车-路连续时变二维行车风险量化建模方法；探究主客观行车风险的认知可信度，揭示人机认知冲突机理，建立考虑认知冲突的驾驶人模型；基于驾驶人的趋利避害决策机制引入最小作用量原理，构建抽象描述驾驶人多目标决策过程的虚拟力学系统，提出基于最小作用量原理的智能汽车人机冲突消解策略。本研究将对智能汽车人机共驾技术的发展及应用提供理论支撑。

The situation that the driver and the intelligent system operate the vehicle together will exist for a long time in developing intelligent vehicle technology. The existing technology can not guarantee the absolute safety of intelligent vehicle driving. The comfort of the human-machine cooperative driving process still needs to be improved, which seriously restricts the development and application of the technology. To this end, this project aims to improve the driving safety and handling comfort of intelligent vehicles as the overall goal. Around the two key scientific issues of "two-dimensional driving risk modeling considering driver's subjective risk perception" and "human-machine cognitive conflict at the level of subjective and objective driving risk assessment," this project deeply studies the driving risk sources and gradient changes. It proposes a two-dimensional quantitative modeling method of driver-vehicle-road continuous time-varying driving risk based on driver's cognitive characteristics; Secondly, it explores the cognitive credibility of subjective and objective driving risks, reveals the mechanism of human-machine cognitive conflict, and establishes a driver model considering cognitive conflict; Finally, we introduce the principle of least action to construct a virtual mechanical system that abstractly describes the driver's multi-objective decision-making process and propose a human-machine conflict resolution strategy based on the principle of least action based on the driver's decision-making mechanism. This research will provide theoretical support for developing and applying human-machine cooperative driving technology of intelligent vehicles.