A Human Motion Prediction Algorithm Based on HSMM for SIAT's Exoskeleton

摘要

Getting the exoskeleton pilot's motion intent is very important for exoskeleton control. In this paper, a new approach to predict human motion is proposed. The prediction method is based on Hidden Semi-Markov Models. The orientation of the pilot's segment is collected by the Inertial Measurement Unit. The Adaptive Boosting algorithm is adopted to classify the orientation data into labels. The sequence of motion labels is inputted into Hidden Semi-Markov Models for motion prediction. The experimental data validate the proposed method.