Hybrid Control of Wheeled Quadruped Robot for Enhanced Mobility and Efficiency in Diverse Terrains

Abstract

In this study, we develop and investigate a control algorithm for wheeled quadruped robots, taking advantage of the combined benefits of legs and wheels. Robots with legs exhibit excellent maneuverability across various terrains but often face speed and energy efficiency limitations. These limitations become critical in activities requiring continuous and dynamic movements, such as disaster relief operations. Conversely, wheeled robots can quickly move across flat terrains but struggle with obstacles larger than their wheel radius. To overcome these challenges, this paper introduces a new wheel-torque control algorithm that enables the robot to maintain a stable posture and move efficiently at desired speeds across different terrains. The algorithm integrates whole-body control through model predictive control and wheel torque optimization using quadratic programming The effectiveness of the developed algorithm is validated through experiments conducted in a Gazebo simulation and with actual hardware. Our method is successfully tested on Pongbot-W, a wheeled quadruped robot developed in our lab. Further videos and results can be found at https://youtu.be/ImV1oqlT1fM.