Stability analysis for passive robots walking on inclined surfaces with local angles

Abstract

For common-legged robots with a kind of stiff joints, the human-like and consecutive gait shown by passive walking robots is a bit difficult to achieve because of varying degrees of control, which is usually accompanied by computational cost. Of course, passive dynamic walking undoubtedly has its inevitable disadvantages because of the lack of control, which is stability. At the same time, the significant nonlinearity of the passive gait increases the difficulty of realizing the walking stability of the passive robot. Therefore, the passive robot still needs certain control to achieve stable walking even on the natural plane or smooth slope. In this paper, a passive robot model walking on an inclined plane with a local angle is studied. First of all, the approximate solution of the nonlinear dynamic equation is given by the perturbation method, and the conditions for the robot to achieve stable walking without external forces are obtained. Further, the input-output feedback linearization control based on hybrid zero dynamics is employed to carry out virtual constraints on the passive robot during the swing stage of walking, facilitating the transition from the unstable state to the stable periodic state. From the result of the experiments, the walking stability of the passive robot is improved over a larger range of walking status values compared with the condition without external force control, and the use of this method reduces the complicated work of solving the fixed point. These findings possibly have reference value in passive walking stabilization control. Meanwhile, the simulation results obtained by designing and studying the minimum controlled walking model show that human walking is based on the uncontrolled mechanical process. The exploration of the mechanism of passive gait can provide some enlightenment to the research of human texture and the application of exoskeleton.