Teoria dos jogos aplicada ao controle de potência e à equalização adaptativa em sistemas de comunicação móvel

Abstract

Game theory is a branch of the Mathematics concerned with the analysis of interactions between competing elements, which are found in conflicting situations, and concerned with the formulation of decision strtegies. This theory is potentially applicable to communications systems problems, since elements in conflicting situations can be identified in some of such problems.Two problems are here considered: the transmit power control and the adaptive channel equalization. Both problems are related to interference, which is one of the most important limiting factors for the cellular system perfomance. Transmit power control consist of a procedure for multi-acess interference management. A new game theoretical approach to power control problem is considered, resulting in a new way to deduce the classical power control algorithm DPC(Distributed Power Con trol). A new algorithm, denoted GT-DPC(Game-theoretic Distributed Power Control), is developed and can be seen as a general form of DPC algorithms for best effort services, since for a unique transmit power level it provides data rates higher than DPC. Furthermore, it allows the power resource management in the presence of services, since for a unique transmit power level it provides data rates higher than DPC. Furthermore, it allows the power resource management in the presence of services with different characteristics. In this case, the algorithm is denoted GT-MSDPC(Game-Theoretic Multi-Service Distributed Power Control). The perfomace of the proposed algorithms for single-service and multi-service systems is demostrated through computational experiments whch simulate TDMA(Time Division Multiple Acess) and CDMA(Code Division Multiple Acess) cellular Systems. The game theory application to adaptive equalization, which is the procedure to combat the intersymbol interference, is related to worst case situations. The H filter(robust filter) is deduced by applying game-theoretic concepts. Furthermore, their relations with the Kalman filter are presented. Through computational experiments wihch simulate GSM(Global System for Mobile Communications) cellular system, both filters have their perfomance as adaptive channel equalizers valued in two different scenarios. In the first one,different speeds are attributed to the user, and results show that both RLS and H equalizer present similar perfomances. In the secon scenario, impulsive noise is considered. Impulsive noise may have external sources, as motors ignition, energy transmission lines or microwaves ovens. In this scenario, the H equalizer robustness is demontrated, so as the RLS perfomance degradation. A hybrid RLS-H equalizer is proposed, obtaining expressive gains with respect to conventional RLS equalizer.