A Wearable Sensor System for Human Motion Analysis and Humanoid Robot Control

Abstract

This paper presents a study on the implementation of a wearable sensor system for quantitative human motion analysis and the application of the captioned sensor system to humanoid robot control. A human motion analysis system based on the wearable sensor system has been constructed, which is considerably inexpensive compared with the conventional 3D motion analysis system based on high-speed camera. Since conventional camera-based system requires costly devices, vast space as well as time-exhausted calibration experiments, the wearable sensor-based system is much low-cost because it is free of those cost factors. The gyroscopes (ENC-05EB) and two-axis accelerometers of ADXL202 are incorporated in the system. The former are taking care of angular velocity measurement of body segments, and the latter are designated to measure accelerations of leg in every single human motion cycle. The current human motion analysis system has been applied in a leg motion (including foot, shank and thigh, etc.) motion analysis. To facilitate the calculation of the gait phases which is fed by the data of sensors output, a fuzzy inference system (FIS) has been developed and incorporated into the sensor system. A digital filter has also been introduced to eliminate the noises from the output of FIS, which enhances the robustness of the system effectively. Experimental studies have been conducted by utilizing the current wearable sensor-based system in a humanoid robot RI control. The result shows the reliability of the system.