A neural network-based model for lower limb continuous estimation against the disturbance of uncertainty

In this paper, a novel prediction model is proposed to estimate human continuous motion intention using fuzzy wavelet neural network (FWNN) and zeroing (ZNN). During walking, seven channel surface electromyography (sEMG) signals data of hip knee are collected, two selected processed from the muscles based on physiological correlation analysis. Then, FWNN built as an recognition model, with sEMG input physical information output. Meanwhile, ZNN exploited into forming hybrid eliminate errors model. Finally, comparative numerical simulations established indicate validity FWNN–ZNN root mean square error (RMSE), absolute (MAE) coefficient determination (R2) evaluation indexes. Results show that can more accurately intention, which lays theoretical foundation for human–robot interaction rehabilitation robots.