Design and Implementation of a Hierarchical Motion Control Framework for Bipedal Robots on Various Structured Terrains

The prevalence of structured terrains in human environments highlights the importance of developing a versatile controller adaptable to multiple terrain types. This paper presents a detailed implementation of a motion control framework that ensures dynamic and robust locomotion on various structured terrains. The framework adopts a hierarchical planning and control architecture, where the high-level module uses the inverted pendulum simplified model without constant height constraint to plan the three-dimensional center of mass (CoM) trajectory through model predictive control (MPC) and the low-level module employs optimization-based whole-body control (WBC) to track and execute the planned trajectories, utilizing the full-body kinematic and dynamic models. This framework is deployed on our self-developed 12-degree-of-freedom (DoF) bipedal robot. Extensive simulation tests demonstrate that the robot can achieve stable and dynamic omni-directional walking on flat ground and exhibit robust performance on other common structured terrains, including stairs and slopes. The results validate the feasibility and effectiveness of the proposed control framework.