Abstract: Metamorphic robot might switch naturally between wheeled driving and foot walking according to the changes of external environments, so that the robot might both run fast on the smooth structure of the road and climb over obstacles in rugged mountains. Based on the generalized coordinate method, a kinematics model of the steering and reconstruction processes of the transformer robots was established. Considering the impacts of the swinging leg in contact with the environment during the reconstruction processes, an impedance control method was proposed to realize the landing smoothness control of the swinging legs. Based on the traditional impedance control, an adaptive impedance controller was designed based on Lyapunov asymptotic stability theorem, and the parameters of impedance control were optimized by particle swarm optimization algorithm. Through simulation analysis under different environmental stiffness, it is proved that the adaptive impedance controller after parameter optimization may follow the expected contact force well, and improve the adaptability of the variable cell robots to the unknown and changeable environments. Finally, a road experiment was carried out for the transformation robot to land on the feet in the reconfiguration processes, which further confirmed the superiority of the optimized adaptive impedance control method.

Key words: metamorphic robot, parking steering reconstruction, kinematics model, adaptive impedance control, particle swarm optimization algorithm

摘要： 变胞机器人能够根据外界环境变化在轮式行驶和足式行走两种运动模式间自然切换，因此兼具在平整结构路面上快速行驶和在崎岖山地越障行走的能力。基于广义坐标法建立了变胞机器人转向重构过程的运动学模型，考虑到重构过程中摆动腿与环境接触时存在较大冲击，提出了利用阻抗控制方法实现摆动腿着地柔顺控制。在传统阻抗控制的基础上，基于李雅普诺夫渐进稳定性定理设计了自适应阻抗控制器，并利用粒子群优化算法对阻抗控制参数进行了优化。通过在不同环境刚度下仿真分析，证实了经过参数优化后的自适应阻抗控制器能够很好地实现对期望接触力的跟随，提高了变胞机器人对未知多变环境的适应性。最后针对变胞机器人转向重构过程中足着地进行了路面实验，进一步证实了优化后的自适应阻抗控制方法的优越性。

关键词: 变胞机器人, 驻车转向重构, 运动学模型, 自适应阻抗控制, 粒子群优化算法