Human-robot interaction for truck platooning using hierarchical dynamic games
Human-robot interaction for truck platooning using hierarchical dynamic games
复制标题
使用分层动态游戏进行卡车队列的人机交互
DOI：
--
复制
发表时间：
2019
期刊：
影响因子：
--
通讯作者：
Karl H. Johansson
中科院分区：
文献类型：
--
作者：
Elis Stefansson;J. Fisac;Dorsa Sadigh;S. S. Sastry;Karl H. Johansson
本文提出了一种用于自动卡车编队的控制器设计框架，以确保与人类驾驶汽车的安全交互。这种交互被建模为一种分层动态博弈，在人类驾驶员和编队中最近的卡车之间进行。分层分解是时间上的，具有预测即时交互的高保真战术层面和估计长时域行为的低保真战略层面。这种分层方法能够进行可行的计算，其中人类的不确定性由量子响应模型表示，并且卡车应使其收益最大化。通过使用驾驶模拟器的案例研究对闭环控制进行了验证，在案例研究中我们将我们的方法与仅使用战术层面的短时域替代方法进行了比较。结果表明，我们的控制器对情况更具感知能力，从而产生自然且安全的交互。
This paper proposes a controller design framework for autonomous truck platoons to ensure safe interaction with a human-driven car. The interaction is modelled as a hierarchical dynamic game, played between the human driver and the nearest truck in the platoon. The hierarchical decomposition is temporal with a high-fidelity tactical horizon predicting immediate interactions and a low-fidelity strategic horizon estimating long-horizon behaviour. The hierarchical approach enables feasible computations where human uncertainties are represented by the quantal response model, and the truck is supposed to maximise its payoff. The closed-loop control is validated via case studies using a driving simulator, where we compare our approach with a short-horizon alternative using only the tactical horizon. The results indicate that our controller is more situation-aware resulting in natural and safe interactions.