汽车驾驶人姿态监测系统研究综述（双语出版）

随着自动驾驶技术的发展，驾驶人将会参与更多的与驾驶无关的活动，从而呈现出新的姿态，这些新姿态是优化传统被动安全系统的重要切入点。而且在未来相当长的时间内，自动驾驶车辆的行驶依然依赖于人和系统的密切配合。对驾驶人姿态的观察，则可以为判断驾驶人是否有能力及时接管车辆提供帮助，从而确保安全、合理的人机交互过程。通过对大量相关文献的系统性梳理，综述了汽车驾驶人姿态监测技术的智能化发展趋势，从传感器种类以及相应的姿态监测算法出发，分析了目前不同监测系统的优缺点。研究发现，尽管传感器技术和姿态识别算法取得了明显进步，然而廉价稳定且能够在实际驾驶条件下对驾驶人姿态准确感知的监测系统依然缺乏。总体而言，目前的监测系统大多只是集中于对驾驶人局部身体部位的感知，缺乏实际驾驶条件下的性能分析，并且对驾驶人状态的实时感知和预测能力仍有待完善。最后，针对目前监测系统所面临的问题，对未来可能的研究方向进行展望，并提出主动式立体视觉系统和压力传感器阵列相融合的驾驶人姿态监测方式。研究成果将为驾驶人姿态监测系统的研究提供参考和借鉴，从而有助于道路交通安全水平的进一步提升，同时也可为人机交互界面的设计带来启发。

Automobile Driver Posture Monitoring Systems: A Review（in English）

With the development of autonomous vehicles, drivers will adopt new postures, which should be considered for optimizing the future passive safety systems. On the other hand, the safety of recent autonomous vehicles is guaranteed by an elaborate interaction between the human driver and the automation system, and this situation will last for quite a long time. Given this context, posture monitoring is necessary to verify the driver's readiness to take over the control when required. Based on a systematic analysis of related literature, a detailed review of recent driver posture monitoring systems classified by the sensor categories and corresponding algorithms was performed. Advantages and disadvantages of different systems were analyzed and summarized. Despite the recent advancements in sensor technologies and algorithms, a cost-effective and robust driver posture monitoring system that can work in various real driving conditions does not exist. In general, most of the existing systems monitor the driver's body parts in isolation for specific research purpose, and few systems consider the body parts in conjunction. Most of the studies were performed in a controlled environment, and lack rigorous and quantitative evaluation in a moving automobile. In addition, real-time posture recognition and prediction methods need to be improved. Finally, future research and development directions to overcome the existing drawbacks are suggested, and a posture monitoring system based on a combination of active stereo vision system and force sensor arrays is recommended. This paper provides valuable insights on developing state-of-the-art driver posture monitoring systems to enhance the road traffic safety. Meanwhile, it also inspires the design of human-machine interfaces.