Legendre Neural Network based Intelligent Control of DC-DC Step Down Converter-PMDC Motor Combination

Abstract

Angular velocity control in DC-DC converter-driven direct current (DC) motors exhibit several challenges in numerous applications. This article proposes a novel single functional layer Legendre neural network integrated adaptive backstepping control technique for the DC-DC step down converter-permanent magnet DC (PMDC) motor system. The proposed controller first aims to estimate the uncertainties in an online mode and then compensate the same efficiently during the robust control action. The closed loop feedback stability of the entire system under the action of proposed controller and the online adaptive learning laws are proved using Lyapunov stability criterion. Further, the proposed controller is numerically simulated for various test conditions including; (a) startup response, (b) a step change in the load torque and (c) reference angular velocity tracking. The transient performance measures of angular velocity such as peak overshoot, peak undershoot and settling time have been observed under the proposed control design and compared with the response obtained from proportional-integral-derivative (PID) controller. Finally, the results presented demonstrate the efficacy of the proposed controller in yielding an enhanced performance under both nominal and perturbed test conditions over a wide operating range.