The Gait Pattern Predicted Model Based on Neural Network

Abstract

Gait pattern planning of the Lower Limb rehabilitation robot, which is the key point of the efficient of rehabilitation. In our work, the walking motion was captured with a webcam using passive markers, and the joint data captured was curve fitted by the Fast Fourier Transform function. For generating the joint angle waveforms of the lower limb during walking, we introduced a method for gait pattern planning. The Generalized Regression Neural Network (GRNN) was applied to predict the Fourier coefficient vectors for a specific subject in this method. After the trained GRNN the Gait Pattern Predicted Model (GPPM) was built. In the end, the quality of the Gait Pattern Predicted Model (GPPM) was examined. The results show that the constructed waveforms closely match the actual waveforms based on the assessment parameter outcomes. From the experiment results, it can be demonstrated that this method is effective.