Lower limb joint angle prediction using square root untraced Kalman neural network based on sEMG signal

The traditional neural network joint angle prediction model has the disadvantages of long training time and difficulty determining the network structure and super parameters. Therefore, the lower limb joint angle prediction using Square Root Untraced Kalman Neural Network (SRUKFNN) based on sEMG signal was proposed. The SRUKF is used to optimize the weight of the neural network to improve the prediction accuracy. The experimental results showed that the overall RMSE of multi-joint angle prediction is 0.78, and the correlation coefficient (R) is 0.99. Compared with other methods, the accuracy of angle estimation was improved by two times.