Climate Control of Greenhouse System Using Neural Predictive Controller

Abstract

This paper presents the concept of neural predictive techniques for the modeling and controlling of the greenhouse system (GHS). Greenhouse system provides the favorable environment to the plants. The GHS is a class of nonlinear and complex systems. Initially, the dynamics of the GHS are precisely modeled in the presence of the uncertainties and disturbances using the system identification approaches based on the neural network (NN). To train the NN, Levenberg–Marquardt backpropagation algorithm is being utilized. This research uses the neural predictive control (NPC) approach to achieve stabilizing control and tracking control. The efficacy of the proposed scheme is validated for the various operating conditions under different initial conditions and enormous external disturbances. The superiority of the proposed research is also compared with the conventional PID control.