Real‐time implementation of adaptive neural backstepping controller for battery‐less solar‐powered PMDC motor

Abstract

Abstract A low‐cost battery‐less solar‐powered PMDC motor using an adaptive backstepping Chebyshev neural network controller to track the desired speed for any change in irradiance and load torque is proposed in this paper. The neural network is used for approximating the variable load torque because of its approximation property. The computational burden of the control law is reduced because of the orthogonal property of Chebyshev polynomials. The asymptotically stable system is obtained by tuning the weights of neurons in accordance with the Lyapunov stability analysis. From the Lyapunov control function of backstepping control design procedure, the control law is obtained by an innovative way of elucidating the cubic equation, in place of resolving the derivative of the control law in the control function. This approach eliminates the constraints caused by the non‐strict feedback system for the backstepping control approach and also this reduces ripples in the duty cycle which makes its appropriateness in real‐time. To ensure its robustness in tracking the desired speed at a faster time and minimum overshoot, simulations are done for an extensive range of variations in irradiance and load torque, and the obtained results are assessed by comparing it with the PID controller and conventional backstepping controller. Because of the use of neural network the robustness of the proposed controller is ensured with enhanced transient and steady state responses. A prototype is developed in the laboratory and the obtained results are assessed by comparing it with the backstepping controller and PID controller.