A real-time gait phase detection method for lower limb exoskeleton based on simple recurrent units

Abstract

A gait phase segmentation method integrating the motion characteristics of lower limb exoskeleton robots was proposed by analyzing their motion data. A neural network based on a simple recurrent unit was also presented for gait phase detection. This method achieves 98.53% accuracy in gait phase detection while meeting the real-time requirements of gait recognition. The proposed method provides a high-precision and real-time control basis for exoskeleton control systems, improving the coordination between the wearer and the exoskeleton robot during movement.