Wang Kui, Wu Xitao, Shi Shaoyang, Xu Mingfan, Han Yifei, Zhu Zhewei, Qin Yechen
School of Mechanical Engineering, Beijing Institute of Technology, Beijing 100081, China.
Sensors (Basel). 2025 May 3;25(9):2888. doi: 10.3390/s25092888.
Human exploration and rescue in unstructured environments including hill terrain and depression terrain are fraught with danger and difficulty, making autonomous vehicles a promising alternative in these areas. In flat terrain, traditional wheeled vehicles demonstrate excellent maneuverability; however, their passability is limited in unstructured terrains due to the constraints of the chassis and drivetrain. Considering the high passability and exploration efficiency, wheel-leg vehicles have garnered increasing attention in recent years. In the automation process of wheel-leg vehicles, planning and mode decisions are crucial components. However, current path planning and mode decision algorithms are mostly designed for wheeled vehicles and cannot determine when to adopt which mode, thus limiting the full exploitation of the multimodal advantages of wheel-leg vehicles. To address this issue, this paper proposes an integrated path planning and mode decision algorithm (IPP-MD) for wheel-leg vehicles in unstructured environments, modeling the mode decision problem using a Markov Decision Process (MDP). The state space, action space, and reward function are innovatively designed to dynamically determine the most suitable mode of progression, fully utilizing the potential of wheel-leg vehicles in autonomous movement. The simulation results show that the proposed method demonstrates significant advantages in terms of fewer mode-switching occurrences compared to existing methods.
在包括山地地形和洼地地形在内的非结构化环境中进行人类探索和救援充满危险和困难，这使得自动驾驶车辆成为这些领域中一种很有前景的替代方案。在平坦地形中，传统轮式车辆展现出出色的机动性；然而，由于底盘和传动系统的限制，它们在非结构化地形中的通过性有限。考虑到高通过性和探索效率，轮腿式车辆近年来受到越来越多的关注。在轮腿式车辆的自动化过程中，规划和模式决策是关键组成部分。然而，当前的路径规划和模式决策算法大多是为轮式车辆设计的，无法确定何时采用哪种模式，从而限制了轮腿式车辆多模式优势的充分发挥。为了解决这个问题，本文提出了一种用于非结构化环境中轮腿式车辆的集成路径规划和模式决策算法（IPP-MD），使用马尔可夫决策过程（MDP）对模式决策问题进行建模。创新性地设计了状态空间、动作空间和奖励函数，以动态确定最合适的行进模式，充分发挥轮腿式车辆在自主移动方面的潜力。仿真结果表明，与现有方法相比，所提方法在模式切换次数更少方面具有显著优势。
