A Prediction Model of Continuous Lower Limb Motion Based on FOX Optimized LSTM

In view of the problems of poor human-computer interaction and unstable movement process in the training process of lower limb exoskeleton rehabilitation robot, a lower limb continuous motion prediction model (FOX-LSTM) based on FOX algorithm (FOX) optimized long and short-term memory (LSTM) is proposed. Based on the surface electromyogram (sEMG), the comprehensive features can well represent the lower limb movement intention, complete signal pretreatment, feature extraction and dimension reduction. And by FOX algorithm for LSTM model optimization, the key parameters of the electromyographic signal characteristics match the joint trajectory, improve the joint Angle prediction accuracy. The experimental results prove that the proposed prediction model can predict the angle information of human lower limb joints well and improve the prediction accuracy.