The Improvement Of Brain Emotion Learning Model And The Application In Neural Network Controller

In robot technology,it is very important to obtain a stable dynamic control method.The control of nonlinear systems is regarded as an important research topic in control engineering.In particular,controllers must well handle both nonlinearly and uncertainty features existing in controlled systems.However,difficulties of mathematical modeling for the nonlinear and uncertain features impede the development of high-performance controllers.Artificial neural networks possess several intelligent features such as selflearning and self-adaptation.Therefore,several artificial neural network-based controllers have been rapidly developed,so as to obtain fast convergence and dynamic response.In the aspect of neural network,this paper proposes three control systems based on brain emotion learning model,In the first network,The proposed neural network is composed of a conventional brain emotional learning network(BEL)and a cerebellar model articulation controller network(CMAC),The structure of the network is dynamic,using a self-organizing algorithm allows increasing or decreasing weight layers;The second network is to integrate the wavelet function into the traditional brain emotion learning model,and introduce the loop structure,recursive mechanism and the advantages of brain emotion learning system,which can optimize the performance of nonlinear system in uncertain environment;The third network integrates the advantages of the first and second networks,applies wavelet network and self-organization mechanism to brain emotional learning network,and changes the original internal circulation mechanism into internal and external double circulation mechanism.The proposed three networks are used to simulate the ideal controller together with the compensation controller,and the parameter updating rules are derived based on Lyapunov stability analysis theory.The effectiveness of the proposed system is verified on different experimental platforms.