Posture prediction of humanoid robot: Modeling and simulation of manual lifting

Abstract

How the central nervous system manage the body posture during various tasks despite redundancy problem? It's a well known question mooted in area of research such as biomechanics and medicine. Some techniques based on muscle and torques synergies presented to express CNS addressing the kinetic redundancy in musculoskeletal system. A 5DOF biomechatronical model of human body subjected to simulate the manual lifting task of humanoid robot. Simulation process is based on optimization approach named predictive dynamics. It uses inverse dynamics to consider the dynamics of motion in simulation process. An objective function in term of ankle torques during lifting time, subjected to be minimized. It assumed that CNS considered this function to perform lifting motion balanced. In the other optimization-based simulations, balancing motion was guaranteed by a nonlinear inequality constraint which restricts the total moment arm of the links to an upper and lower boundary. In this method there is no need to use this constraint. Result shows that the motion is performed balanced. According to the comparison the results with the experimental data, the body posture of humanoid robots, predicted as similar as actual human posture.