Hybrid intelligent control for set-point tracking of a flexible manoeuvring system using sigmoid activation function

Abstract

This article addresses the input tracking control problem of a highly nonlinear flexible system namely a twin-rotor multi-input multi-output system in the vertical plane motion. The twin-rotor multi-input multi-output system can be considered as an aerodynamic experimental model, representing the control challenges of complex air vehicles. A hybrid control scheme comprising a fuzzy proportional–derivative-like control and a conventional integrator is developed by adding the outputs of both control actions. To satisfy the tracking characteristics of the manoeuvring system, an efficient combination of the control actions is developed using a smooth nonlinear activation function which continuously regulates contribution of the integral term in the controller output. The effect of the activation function greatly reduces overshoots when the system is exposed to abrupt changes during its manoeuvre. The control system performance is tested in simulations and experimentally in real-time situations. Robustness and stability of the controller are also investigated. The obtained results prove the efficiency of the proposed scheme in terms of time-domain performance specifications of the system.