Comparison of force estimation using surface EMG from forearm muscles

The surface electromyography (sEMG), which is a biological signal, has been applied in many fields related to human. However, there have been few studies on the relationship between the force of agonist and antagonist muscles to estimate handgrip force from sEMG signals. The fast and precise estimate is required in a real-time application. In this study, extreme learning machine (ELM) is applied to estimate force using sEMG signals from flexor carpi radialis (FCR) and extensor digitorum (ED) muscles in the forearm. The system is trained with a root mean square feature computed using the sEMG data from a handgrip action. The root mean square feature was taken as the input of ELM and the force value was estimated from the ELM output. Experimental results showed that the estimations of forces from FCR and ED muscles give the mean and standard deviation of root mean square errors at 1.4067 ± 0.3740 kg and 1.6869 ± 0.6580 kg, respectively. Results from the error values show that the proposed method can be properly applied to estimate force. Moreover, there is no significant difference for force estimate using the sEMG data from FCR and ED muscles.