A fuzzy logic based neural network controller for the flexible pole-cart balancing problem
Date of Publication
Master of Science in Computer Science
Computer and Systems Architecture | Controls and Control Theory | Digital Communications and Networking | Hardware Systems | Systems and Communications
College of Computer Studies
Elmer P. Dadios
Defense Panel Chair
Arnulfo P. Azcarraga
Defense Panel Member
This paper investigates the feasibility of developing a controller based on a fuzzy reasoning that is implemented by an artificial neural network (NN) to control complex and highly nonlinear systems. The flexible pole-cart balancing problem (FPCBP) is used as a test bed for this application as it was shown by a study to exhibit more nonlinearity as compared to the conventional inverted pendulum problem. The neural networks are used as a tool in determining the appropriate number of rules, in determining the fuzzy inference rules, i.e., learning and generating the membership functions (antecedents) and the functions that represent the consequents, and finally, in implementing the fuzzy logic based controller. The use of this method in building the controller eliminates extensive knowledge from the human expert. An off-line fuzzy logic based NN software controller for the flexible pole-cart balancing system (FPCBS) is developed using a set of training data taken from the results of a real time computer simulation of the FPCBS derived dynamics. Results show that the controller developed is able to predict the desired control outputs with high accuracy.
Archives, The Learning Commons, 12F Henry Sy Sr. Hall
xi, 79 leaves; 28 cm.
Logic machines; Fuzzy systems; Neural networks (Computer science); Computer network; Electric controllers; Sequence controllers; Programmable
Gunay, N. S. (1998). A fuzzy logic based neural network controller for the flexible pole-cart balancing problem. Retrieved from https://animorepository.dlsu.edu.ph/etd_masteral/1983