Variable Horizon MPC With Swing Foot Dynamics for Bipedal Walking Control

In this letter, we present a novel two-level variable Horizon Model Predictive Control (VH-MPC) framework for bipedal locomotion. framework, the higher level computes landing location and timing (horizon length) of swing foot to stabilize unstable part center mass (CoM) dynamics, using feedback from CoM states. The lower takes into account dynamics generates dynamically consistent trajectories at desired time as close possible location. To do that, use simplified model robot projected in space that joint torque constraints well friction cone stance foot. We show effectiveness our proposed control by implementing robust walking patterns on torque-controlled open-source biped robot, Bolt. report extensive simulations real experiments presence various disturbances uncertainties.