当前位置: X-MOL 学术Proc. Inst. Mech. Eng. Part D J. Automob. Eng. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
A novel evaluation method for vehicle and traffic performance of different decision control of automatic lane change based on miniature model
Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering ( IF 1.5 ) Pub Date : 2020-05-21 , DOI: 10.1177/0954407020919578
Yangyang Wang 1 , Guangda Chen 2 , Yuanxing Jiang 1
Affiliation  

Research on automatic lane-change decision is mainly limited to simulation validation and lacks real vehicle validation methods because it is limited by experimental site and automatic driving technology on real vehicles. This paper puts forward a miniature traffic model to simulate the actual traffic scene and achieves to verify the decision control of automatic lane-change scene. The miniature intelligent traffic scene contains miniature vehicles, simplified miniature road traffic environment, and wireless network communication. After testing the basic functions of the miniature traffic scene model, such as automatic lane change, lane keeping, and automatic following, a semi-physical simulation test of the traffic flow composed of the model vehicle and the virtual vehicle is carried out. The semi-physical simulation test includes vehicle-following test of hybrid-condition intelligent driver model, lane-change test of lane-change decision two-vehicle gaming model, and minimizing overall braking induced by lane changes. The results show that the feasibility of the method and of the lane-change decision two-vehicle gaming model of automatic lane change is better in terms of traffic safety, traffic efficiency, and homogeneity. Compared to the minimizing overall braking induced by lane-change model test, the test of lane-change decision two-vehicle gaming model improves 2.26% and 1.5% in the average speed and total driving distance, respectively. The standard deviation of the traffic speed of the lane-change decision two-vehicle gaming model was 28.57% lower than the minimizing overall braking induced by lane changes. Compared to pure simulation verification, the method considers the effects of actual sensor signals and actuator control, which is closer to the actual application.

中文翻译:

基于微型模型的自动换道不同决策控制下车辆及交通性能评价新方法

自动变道决策的研究主要局限于仿真验证,缺乏实车验证方法,因为受实验场地和实车自动驾驶技术的限制。提出了一种微型交通模型来模拟实际交通场景,实现对自动换道场景决策控制的验证。微型智能交通场景包含微型车辆、简化的微型道路交通环境和无线网络通信。在测试了微型交通场景模型的自动变道、车道保持、自动跟车等基本功能后,对模型车和虚拟车组成的交通流进行了半实物仿真试验。半实物仿真测试包括混合工况智能驾驶模型的跟驰测试、变道决策两车博弈模型的变道测试,以及最小化变道引起的整体制动。结果表明,该方法与自动变道的变道决策两车博弈模型在行车安全性、行车效率和同质性方面的可行性较好。与最小化换道模型测试引起的整体制动相比,换道决策两车博弈模型的测试在平均速度和总行驶距离上分别提高了 2.26% 和 1.5%。变道决策两车博弈模型的交通速度标准偏差比最小化变道引起的整体制动低28.57%。
更新日期:2020-05-21
down
wechat
bug