Weighted differential evolution-based heuristic computing for identification of Hammerstein systems in electrically stimulated muscle modeling

In this paper, weighted differential evolution (WDE)-based evolutionary heuristic computing is presented for parameter estimation of nonlinear dynamics of the electrically stimulated muscle system based on a Hammerstein model structure as an element of rehabilitation program for the stroke patient. The approximation theory is defined on the basis of mean square error of actual and estimated responses for the system's fitness function. Optimization technique based on WDE is employed to find the optimal parameters of ESM model with sigmoidal, polynomial and spline as input nonlinear functions. Statistical analysis verified the worth of WDE as computational heuristic technique for identification of electrically stimulated muscle system.

Automated summary

This paper presents a weighted differential evolution (WDE)-based evolutionary heuristic computing method for parameter estimation of the nonlinear dynamics of the electrically stimulated muscle system. The fitness function is defined based on the mean square error between the actual and estimated responses. The optimization technique using WDE is applied to find the optimal parameters for the ESM model, with different input nonlinear functions. Statistical analysis confirms the effectiveness of WDE as a computational heuristic technique for identifying the electrically stimulated muscle system.