Electrical Muscle Stimulation Models Identification Based on Hammerstein Structure and Gravitational Search Algorithm

In this study, the electrical muscle stimulation (EMS) models are effectively estimated by representing them in Hammerstein structure form. Moreover, a highly computational efficient population-based evolutionary optimisation algorithm (EOA) named as gravitational search algorithm (GSA) is employed to get the optimal coefficients of the unknown EMS systems. Due to exploration and exploitation phases, GSA is able to avoid the local stagnation problem, unlike the genetic algorithm (GA) and fully informed particle swarm optimization (FIPSO). In this paper, three different EMS plants having different nonlinearities such as polynomial, sigmoid, and cubic spline, respectively, are successfully identified by using the real coded GA (RGA), FIPSO, and the GSA techniques. The simulation results confirm that the GSA exhibits more robust performance and accurate identification results as compared with the RGA and FIPSO methods which have been verified by using various quantitative metrics.