Impact Force Reduction Using Variable Stiffness with an Optimal Approach for Falling Robots

The work described in this paper is focused on the reduction of the impact force exerted by the ground on a falling humanoid robot. It proposes the use of a variable stiffness in the arms's motors to prevent damages. The proposed work is applicable when the falling prevention techniques fail or when falling is unavoidable. This work proposes the generation of variable stiffness in a motor through the optimal design of a PID controller. The variation of the Q matrix in a LQR controller produces different levels of motor stiffness. The proposed variable stiffness is tested in a Darwin OP Robot. The performance of the proposed design is evaluated using the estimation of the impact force. Results show an impact force reduction on falling motions by means of stiffness variation.