虚幻4驾驶模拟器设计和驾驶员特性分析

深入研究人类驾驶员的驾驶行为和特性,对推动智能汽车向高度自动驾驶发展具有重要意义。采用驾驶模拟器研究驾驶员在复杂交通场景下的驾驶决策和驾驶行为,成为近年来的一个研究热点。基于虚幻四引擎UE4交互式视景仿真技术,通过车辆、道路、建筑、交通灯、行人、路牌等交通元素的驾驶视景仿真环境搭建,开发了CarSim汽车动力学模型和罗技G29力反馈方向盘踏板的具有高拟真度人机交互的驾驶模拟系统。并设计了典型工况下的驾驶模拟试验,通过实时采集驾驶员驾驶数据,对驾驶员特性进行研究。研究结果表明:该驾驶模拟器具有逼真的驾驶体验,利用模糊C聚类算法(FCM,fuzzy C-means),将驾驶员特性进行准确分为6个聚类,可以将驾驶员特性进行准确分类,确立驾驶员特性与驾驶能力的关联,为进一步建立实时驾驶权分配研究奠定了基础。

Research on driver characteristics based on unreal engine 4 driving simulator

In-depth study of the driving behavior and characteristics of drivers is of great significance in promoting the development of intelligent vehicle to highly automatic drive. The application of driving simulator to the study of the driver''s driving decision and driving behavior in complex traffic scenarios has become a research hotspot in recent years. Based on the unreal engine 4(UE4) interactive scene simulation technology, we built a driving simulation environment with many traffic elements such as vehicles, roads, buildings, traffic lights, pedestrians, road signs, etc. and designed the high-fidelity and human-machine interactive driving simulation system of Logitech G29 force feedback steering wheel pedal set with CarSim vehicle dynamics model. With the driving simulation system, driving simulation experiments were conducted under typical traffic scenarios and driver''s information and real-time driving data were collected to study driving behavior and characteristics. The results show that by the driving simulator a realistic driving experience can be got. Based on Fuzzy C-means clustering algorithm (FCM), the driver characteristics are classified accurately into six clusters, the driver behavior and driving style can be analyzed, and thus the association between driver characteristics and driving capacity are established. An effective way is provided for further research on the allocation of driving rights concerning real-time driving ability.