Electrically Stimulated Lower Limb using a Takagi-Sugeno Fuzzy Model and Robust Switched Controller Subject to Actuator Saturation and Fault under Nonideal Conditions

Abstract

Electrically stimulated lower limb systems contain higher order nonlinearities and uncertainties in their physical parameters. Takagi-Sugeno (TS) fuzzy models are used to model nonlinear systems. Techniques such as parallel distributed compensation (PDC) are dependent on the membership functions that constitute the TS fuzzy model. When the exact representation approach is used to electrical stimulation applications, the system’s performance under PDC control can be deteriorated, because the membership functions may be uncertain, besides a high computational cost be required to compute them. In this paper, we propose a robust switched control subject to actuator saturation and fault (RSwASF) that effectively handles system uncertainties and nonidealities, such as fatigue, spasms, tremor, and muscle recruitment. Control techniques based on TS fuzzy modeling (PDC and robust PDC), as well as other approaches, such as sliding-mode control, backstepping, super-twisting, gain-scheduling, and proportional-integral-derivative (PID) control were compared to RSwASF through the root-mean-squared error (RMSE). The results indicate that RSwASF minimizes the influence of the parametric uncertainties and presents the lowest RMSE for healthy and paraplegic individuals.