Fuzzy Inference Systems Applied to the Analysis of Machine Mechanisms

D. Sprečić, T. Konjić, E. Mujić (Bosnia-Herzegovina), and V. Miranda (Portugal)


Neuro-Fuzzy Inference Systems, Computational Intelligence, Computer Aided Design


Changing only one of the input variables within a certain range, the analyzing methods (graphic ones) used so far for kinematics and dynamic analysis of mechanisms become unacceptable for a description. If we increase the number of the input variables this problem becomes more complex. Therefore, the use of computer-based methods in solving such problems is a logical answer. The database, which consists of the software calculation results of kinematics and dynamic values of the observed mechanism, is used for analysis of dependence of input output variables. In this paper an Adaptive Neuro-Fuzzy Inference System (ANFIS) was used with purpose of solving problems of mechanism analysis. An efficient method of training the FIS is presented and its quality verified by comparing its performance with analytical software tools.

Important Links:

Go Back