Four Wheeled Humanoid Second-Order Cascade Control of Holonomic Trajectories

Abstract

This work develops model-based second-order cascade motion controller of a holonomic humanoid-like wheeled robot. The locomotion structure is comprised of four mecanum wheels radially arranged. The model is given as a function of all wheels contribution adding maneuverability to upper limbs. High-order derivatives are synchronized through numeric derivations and integration, obtained online for consistent performance of inner loops feedback. The controller deploys reference inputs vectors, both global and local to each cascade loop. In this approach, the controller decreases errors in position, velocity and acceleration simultaneously through Newton-based recursive numerical approximations. A main advantage of this approach is the robustness obtained by three recursive feedback cascades: distance, velocity and acceleration. Observers are modeled by combining multisensor inputs. The controller showed relative complexity, effectiveness, and robustness. The proposed approach demonstrated good performance, re-routing flexibility and maneuverability through numerical simulations.