A new control strategy based on an novel improved credit assigned CMAC inverse model implemented in Artificial Climate Chamber network

Artificial Climate Chamber (ACC) is widely used in agricultural research as a simulation of the natural environment of the laboratory equipment. The real-time control of the ACC system is very difficult, due to its serious coupling, hysteresis and nonlinearity. In the paper, an intelligent real-time control strategy based on a novel Improved Credit Assigned Cerebellar Model Arculation Controller Neural network (ICA-CMAC) inverse model is proposed. The novel ICA-CMAC algorithm can not only accelerate the network´s convergence rate, but also reduce interference in learning. Additionally, the inverse model is designed to accurately replicate the inverse system behavior, easy to be applied and has clear physical meaning. Simulation results prove that such a new type of intelligent control strategy is able to effectively achieve fast, stable, real-time tracking control of ACC.