B36 加速度センサとジャイロセンサを用いた歩行周期検出(生体計測・解析1)

Abstract

Objective. To assess the potential use of accelerometer and gyroscope as a sensor for the control of FES assisted walking for the correction of drop foot in hemiplegic individuals. A machine learning technique (Neural Network) was used for the detection of swing phase of the gait. Methods. Two subjects (able-bodied and hemiplegic individual) wore accelerometer and gyroscope over their anterior proximal shank. Footswitch placed on the sole of one foot recorded the heel contact and heel off times for that foot. For Neural Network training, acceleration and gyro data were processed with the input data, and the heel data was processed with the target data. The microcomputer produced output signals using the Neural Network program. The accuracy of the Neural Network detector was compared with a swing phase detector based on the heel sensor. Results. The largest difference in timing of the swing phase was less than 0.03 sec in normal subject and 0.02 in hemiplegic patient. Conclusion. The Neural Network detector could correspondingly detect the swing phase of gait. The present system has a potential to access the reconstruction of FES assisted walking in hemiplegic individuals.