User Centric Interference Management in Wireless Networks (UCIMa)

点击翻译按钮获取中文摘要

In order to serve the growing number of wireless devices and their demand for increased data rates, heterogeneous dense wireless networks will be deployed. Due to the coexistence, one challenge is interference. In the final part of the UCIMa project we focus on one instance of a heterogeneous network namely the hybrid access femto cell network. Many operators have launched femtocell service, including Vodafone, SFR, AT&T, Sprint Nextel, Verizon and Mobile TeleSystems. The femtocell access points (FAP), also known as home base station (BS), are connected to the operators' macrocell networks by backhaul digital subscriber line (DSL), optical fiber or other high speed connections.From the theoretical research of the first two phases in the UCIMa project, we will apply the results to the hybrid access framework of femtocell networks. In this phase, the project is intended to fundamentally investigate the economic framework of the hybrid access between the macrocell and the femtocells. The market model for the physical layer resource allocation for the two-tier hybrid access is going to be established.Jamming is a concern in femtocell networks with hybrid access, since the femtocell users and the unregistered macrocell users served by FAP compete for the resources. In addition, the macro users served by the FAP could create jamming to the other macrocell users under MBS without any risk. From the game theoretic point of view, the users have incentives to cheat on their private information, e.g. the channel state information (CSI) for better utilities. We study the problem of jamming and cheating in the hybrid access mode where users migrate freely between MBS and FAP.We extend the framework in phase two to analyze convergence and decentralized learning for the case of wireless network optimization in the presence of adversarial users. The dynamics and convergence properties of an iterative distributed pricing algorithm with malicious users are analyzed. Next, the assumption in phase two that the malicious user has information about the utility function of selfish users is relaxed and the scenario where the malicious user applies a regression-based iterative learning scheme is analyzed. We propose a Bayesian analysis extension to the work carried out in WP3 in phase two taking into account the transmission energy costs of theusers. In addition, we aim to obtain the most general scalar strategy efficient mechanisms for interference coupled systems similar in spirit of VCG-Kelly mechanism.The competitive market among FAPs have a lot of potential to provide cheap and better service for the end users. We extend the single FAP scenario studied in phase two to a scenario where the multiple FAPs compete for the Macro users to be served by them. In addition, we study the case where some of the FAPs form coalitions and jointly enjoy the benefit.