Synthesizing a desired trajectory and sensory feedback control laws for an origami-folding robot based on the statistical characteristics of direct teaching by a human

摘要

In this paper, a novel method to synthesize a desired trajectory and sensory feedback control laws for robots based on the statistical characteristics of direct teaching data by a human is proposed. This work was motivated by a poor performance of an origami-folding robot developed by the authors. Since the robot simply replayed a given trajectory without sensory feedback control, it often failed in folding due to the fluctuation of origami paper behaviors. To model the statistical characteristics of the demonstrated motions by a human, a hidden Markov model (HMM) is employed. A nominal desired trajectory is obtained by temporally normalizing and spatially averaging the demonstrated motions in a statistical manner. Sensory feedback control laws are then synthesized based on the output probability density function parameters of the HMM. Considering the velocity variance and the canonical correlation between velocity and force of the teaching data, important motion segments are extracted and the feedback control is applied only for those segments. The proposed method was applied to the origami-folding robot and experimental results showed that the success rate and the folding quality of "Valley-fold" were improved. Although the demonstrated task is very specific, the proposed method has generality to be applied to other tasks.