Whole-Body Control Loco-manipulation strategy for quadruped robots on deformable terrains

This paper introduces a loco-manipulation strategy for a quadruped robot operating on deformable terrains. A complete spatial robot model is built as the controlled system. Linear and nonlinear spring-damper models are adopted to represent the terrain deformation. The Operational Space Formulation (OSF) is employed to map the configuration space dynamics to task space dynamics. The Model Predictive Control (MPC) methodology is studied and implemented to direct the controlled system. The operational space motion equations and their interaction with deformable terrain are used as the fundamental model for the MPC framework. A force control strategy is employed on the legs to counteract gravitational effects and stabilize the body on deformable terrain, while a motion control strategy is applied to the arm for the object-manipulating task. The simulation results demonstrate that the proposed controller is feasible for these applications.