Humanoid robot motion generation scheme for tasks utilizing impulsive force

In order to exert a large force on an environment, it is effective to apply impulsive force. We describe the motions in which tasks are performed by applying impulsive force as "impact motions." This paper proposes a way to generate impact motions for humanoid robots to exert a large force and the feedback control method for driving a nail robustly. The impact motion is optimized based on a three dimensional model using sequential quadratic programming (SQP). In this research, a nailing task is taken as an example of impact motion. A dominant parameter for driving a nail strongly is revealed and motions which maximize the parameter are generated considering the robot's postural stability. In order to evaluate the proposed scheme, a life-sized humanoid robot drives nails into a plate made of chemical wood. The optimized motion is compared with a motion designed heuristically by a human. Average driving depth is clearly increased by the proposed method.