Abstract: The transition stage from manual driving to autonomous driving in human-machine driving will exist for a long time and become one of the difficulties that need to be broken through in the development of intelligent vehicles. In order to further study human-machine driving, the definition, classification and research methods of human-machine driving at home and abroad was combed. The existing problems of vehicle in single vehicle intelligent human-machine driving was summarized, and the solution was put forward. It was shown that the main problems at the perception and decision levels were the limited range of environmental perception and the local optimization of decision results. Then, in view of the rise of vehicle infrastructure cooperative technology in recent years, based on the idea of technology change that broke the limitation of single point thinking with systematic thinking, the new method brought by the idea for the enhancement of the perception of human-machine driving vehicles and the collaborative decision-making of driving modes was analyzed. Finally, according to the participation degree of “human-vehicle-road-cloud” traffic elements, the development stages of human-machine driving were divided; meanwhile, a preliminary solution to the problems existing in the current research on human-machine driving of intelligent vehicles was proposed.

Key words: vehicle engineering; intelligent vehicle; intelligent transportation; autonomous driving; vehicle infrastructure cooperative; human-machine driving

摘要： “人机共驾”人工驾驶到无人驾驶的过渡阶段在一段时间内将长期存在并成为智能汽车发展需要突破的难题之一。为进一步研究人机共驾，首先，梳理了国内外有关人机共驾的定义、分类和研究方法，总结了单车智能化人机共驾车辆存在的问题，并提出了解决方案。研究结果表明，感知与决策层面存在的环境感知范围受限、决策结果局部最优是主要问题。然后，针对近年来兴起的车路协同技术，借鉴以系统性思维打破单点思维局限性的技术更迭思路，分析其对人机共驾车辆感知增强及驾驶模式协同决策带来的新方法。最后，根据“人-车-路-云”交通要素的参与程度对人机共驾的发展阶段进行了划分，同时对当前智能汽车的人机共驾研究存在的问题提出了初步的解决思路。

关键词: 车辆工程；智能汽车；智能交通；自动驾驶；车路协同；人机共驾