arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2112.02041eess.SYcs.SY
收藏

混合交通中自动驾驶汽车的社会兼容控制设计

Socially Compatible Control Design of Automated Vehicle in Mixed Traffic

Mehmet Fatih Ozkan, Yao Ma

摘要:

在车辆跟驰场景中,自动驾驶车辆(AV)通常规划运动而不考虑其动作对后续人类驾驶员的影响。本文旨在利用这些影响,以规划更有效的和社会所需的AV行为在人与AV的互动。具体来说,我们引入了一个社会兼容的控制设计的AV,有利于混合交通的汽车以下的情况。该设计通过将心理学中的社会价值取向整合到其决策过程中,使人-AV交互中的利他AV成为可能。利他AV通过优化其自身的奖励和对纵向运动中的以下人类驾驶员的原始计划的礼貌来生成社会期望的行为。结果表明,与利己自动驾驶相比,利他自动驾驶显著地避免了对后车驾驶员初始计划的干扰,并使后车驾驶员的跟车间隔距离和车头时距显著减小.此外,我们研究了社会兼容的控制设计与不同的利他主义水平的AV使用统计评估的影响。结果共同证明了显着改善交通水平的指标作为一个结果的AV的利他行为在人类AV的互动。

英文摘要:

In the car-following scenarios, automated vehicles (AVs) usually plan motions without considering the impacts of their actions on the following human drivers. This paper aims to leverage such impacts to plan more efficient and socially desirable AV behaviors in human-AV interactions. Specifically, we introduce a socially compatible control design for the AV that benefits mixed traffic in the car-following scenarios. The proposed design enables the altruistic AV in human-AV interaction by integrating the social value orientation from psychology into its decision-making process. The altruistic AV generates socially desirable behaviors by optimizing both its own reward and courtesy to the following human driver's original plan in the longitudinal motion. The results show that as compared to the egoistic AV, the altruistic AV significantly avoids disrupting the following human driver's initial plan and leads the following human driver to achieve considerably smaller car-following gap distance and time headway. Moreover, we investigated the impacts of the socially compatible control design with different altruism levels of the AV using statistical assessments. The results collectively demonstrate the significant improvement in traffic-level metrics as a result of the AV's altruistic behaviors in human-AV interactions.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑