A self‐stabilised walking gait for humanoid robots based on the essential model with internal states

Abstract Walking stability is one of the key issues for humanoid robots. A self‐stabilised walking gait for a full dynamic model of humanoid robots is proposed. For simplified models, that is, the linear inverted pendulum model and variable‐length inverted pendulum model, self‐stabilisation of walking gait can be obtained if virtual constraints are properly defined. This result is extended to the full dynamic model of humanoid robots by using an essential dynamic model, which is developed based on the zero dynamics concept. With the proposed method, a robust stable walking for a humanoid robot is achieved by adjusting the step timing and landing position of the swing foot automatically, following its intrinsic dynamic characteristics. This exempts the robot from the time‐consuming high‐level control approaches, especially when a full dynamic model is applied. How different walking patterns/features (i.e., the swing foot motion, the vertical centre of mass motion, the switching manifold configuration, etc.) affect the stability of the walking gait is analysed. Simulations are conducted on robots Romeo and TALOS to support the results.