Human-Riding Inspired Acceleration Control of a Wheel-Legged Humanoid Robot

Wheel-based robots and leg-based robots are two major types of ground mobile robots. It is well known that their performance properties are complementary. With the combination of the two, several of wheel-legged humanoid robots (WLR) are proposed to overcome obstacles on broken pavement and move at high speed on flat pavement. However, WLR's high-speed performance, particularly, during acceleration phase is still not good enough compared with traditional wheeled mobile robots. To be detailed, WLR always response slowly and get large velocity tracking error under the influence of its specific dynamic characteristics. In this case, this paper analyses the conflict between the system's nonlinearity and its inertial dynamics of WLR. Besides, to solve the problem, we present a human-riding inspired acceleration control (HIAC) of a wheel-legged humanoid robot. HIAC contains a partial feedback linearization controller and a human-riding inspired body swing motion. During acceleration phase, it can response quickly and smoothly, also with less error. We demonstrate HIAC's superiority in acceleration by simulation.