In this paper the structures of Adaptive Neuro-fuzzy interface system (ANFIS) are studied for noise identification. The system's structures are analyzed for different types of membership functions applied for input variables with root mean square errors variation. Hybrid algorithm and back propagation algorithm are applied. The input data are obtained through system simulation based on LabVIEW system design platform and development environment. The choice of ANFIS structure is based on the training results and minimum RMSE for identification of the signals with uniform white and inverse F Noises. Therefore, "gbellmf" membership function for input data variables is chosen. The accuracy classification is obtained at 100 %.

In this article they are investigated the problems associated with the measuring of user load in the network versus time and type of equipment. Among these they are presented also practical researches concerning the behaviour of various analytical models used to describe the interaction between the base and mobile stations. The aim is to make proper planning of the network for providing quality services and the necessary traffic in it.