Research On Improvement And Optimization Of Fuzzy CMAC Neural Network And Its Application In Robotics

Dynamic control,including dynamic control of robots,faces both the theoretical challenge of obtaining accurate system models and the practical difficulty of defining fuzzy system boundary.Conventional methods either rely on establishing accurate models or adopting adaptive methods for on-line tuning;however,the control effects of these methods need to be improved.To facilitate such challenges,a new kind of fuzzy neural network is established to obtain better nonlinear characteristics,and a new type of fuzzy neural network and robust compensation control system is also proposed.This control system can make full use of the characteristics of the established neural network model.Then,in view of the optimization method of neural networks,a new harmony search algorithm is adopted to optimize the weights.Finally,on the actual robot system,a double neural network method is adopted to realize the control of robot eye-hand coordination.In terms of neural networks,this paper integrates a number of key components embedded in a Type-2 fuzzy Cerebellar Model Articulation Controller(CMAC)and a brain emotional learning controller network(BELC),thereby mimicking an idealized sliding mode controller.The system inputs are fed into the neural network through a Type-2 fuzzy inference system(T2FIS),with the results subsequently piped into sensory and emotional channels which jointly produce the final output of the network.That is,the proposed network estimates the non-linear equations representing the idealized sliding model controllers using a robust compensator controller with the support of T2FIS and BELC,guaranteeing robust tracking of the dynamics of the controlled systems.The adaptive dynamic tuning laws of the network are based on the exploitation of the popular brain emotional learning rule and the Lyapunov function.In terms of the optimization,the paper studies the problem that the weights are difficult to be determined in CMAC.The paper finds that the harmonic search algorithm shows faster search speed,performance and simple structure when dealing with the optimization problem.It is very suitable for the optimization of network parameters and weights of CMAC.The proposed system is suitable for robot arm,and a double neural network control system is established by simulating the characteristics of human grasping.The experimental comparison with traditional neural network shows that the proposed system has significant improvements in the implementation of Intelligent D-range Control.