Automatic generation of optimal standing-up motion using nonlinear receding horizon control

Care work such as lifting a cared person is a heavy task. To reduce the hard labor of this kind, various power assist technologies are investigated. Their main focus is on the proposal of mechanism or robotic system rather than the optimality of the dynamical behaviour. In this paper, we consider the question, 'what is the best way to make a person stand up?' from the viewpoint of systems control. After deriving the model of the human body as a system of links, the torque minimization problem with terminal condition is considered. To prevent the falling over, the constraint on the Zero Moment Point (ZMP) is introduced. We employ the nonlinear receding horizon control to solve this nonlinear optimization problem with constraints. The generated motion is verified on a multibody dynamics simulator. Finally, the behaviour of the receding and infinite horizon settings is compared through an illustrative example.