Detection Of Shank Movement Based On Empirical Mode Decomposition of EMG Signal and Hidden Markov Model

Exoskeletons are tools for helping specific patients with spinal cord injury, elderly people and even ordinary people. The most important issue in designing these tools is the ability to determine the motor behavior and body position of the person who is using them accurately. In this paper, we have tried to detect the lower body position in semi-sitting mode and standing mode using the electromyogram signal processing in the time-frequency domain and the use of empirical mode decomposition techniques (EMD) as well as using the Hidden Markov Model. Contrary to conventional methods in choosing the Baum-Welch algorithm to determine the parameters of Hidden Markov Model, in this paper we used Genetic Algorithm to estimate the optimal model parameters. In this paper, 5 subjects were used in the age range of 25 years, all of them were right-handed and had complete behavioral and motor health. Feature vectors from three subjects for training and extracted features vectors from two other subjects for testing were blindly given to the estimator model. The results showed that the proposed method with an accuracy of 92% compared to other studies has more accuracy and also less computation compared to other studies.

کلیدواژه ها:

determining lower body position ، empirical decomposition mode ، genetic algorithm ، hidden Markov model ، muscle electromyogram