Model predictive control for biped robots in climbing stairs

Abstract

In this paper, nonlinear model predictive control method is used for biped robots in climbing stairs. The main feature of the proposed approach is that it does not need any offline or on-line trajectory planning. The up-stair gait is formulated such that the environmental constraints and the stability constraints of the robot are satisfied. Another advantage of the proposed method is that the robot is able to climb stairs without knowing the width and height of the stairs. The proposed method is applied to a 5-DOF biped robot in sagittal plane. The effectiveness of the method is shown through simulations.