Sigmoid Function Based Behavioral Model for Standing and Sitting

Abstract

This paper introduces a new model composed of two computational functions describing standing-and-sitting behaviors toward the development of a standing/sitting assist robot. In this model, the one function estimates the locations of hip joints by employing relative distances. The other linearizes the behavioral trajectories estimated by the first function. Specifically, through the linearization, the standing and sitting behaviors are generalized as a well-known sigmoid function. As a result, the proposed model allows to obtain the geometrically linearized representation of the standing and sitting behaviors. This paper details the description of the proposed model. And, through extensive experiments, the effectiveness of the model is verified.