Authors : Seyed Mohammad Reza Faritus and Hadi Homaei
Abstract: In this study, trajectory tracking control of planar serial robots with the last flexible arm is studied. The EOMs are derived using Lagrangian mechanics and the assumed modes method. The robot has a fuzzy controller with neural tuner. The control system consists of a fuzzy logic controller in the feedback configuration with error and change in error of the joints angular as input variables. Set parameters of membership functions of inputs in fuzzification and output in defuzzification are fuzzy control challenges. Utilizing a three-layer perceptron neural network a new method to estimate the on-line self-tuning parameters of the membership functions is presented. In this method, symmetric triangular membership functions are used in fuzzification and defuzzification units. Each of them is function of a productive parameter which is calculated on-line using neural network. The network inputs depending on which membership function is set, can be deformation, joint angular error or its derivative. The back propagation learning algorithm is used to update the network weights and the biases. To validate the proposed method simulation is done and the results are investigated.
Seyed Mohammad Reza Faritus and Hadi Homaei, 2014. Fuzzy Control of Flexible Serial Robots with Neural Tuner. International Journal of Signal System Control and Engineering Application, 7: 61-69.