A generalized cooperative and cognitive radio resource management architecture for future wireless networks

Heterogeneity and convergence are two distinctive connotations of future wireless networks. Multiple access networks are expected to converge in a manner where heterogeneity can be exploited as an enabler to realize the Optimally Connected, Anywhere, Anytime vision of the International Telecommunication Union (ITU). This stimulates current trends toward the convergence of complementary heterogeneous access networks in an all-internet protocol (IP) core network and raises the importance of cooperation in such a multiple radio access technology (multi-RAT) environment. This thesis defines, develops, implements, and analyzes a novel generalized cooperative and cognitive RRM (CCRRM) architecture, anchored on the key principle of technology agnostic approach, to optimize radio resources usage, maximize system capacity, and improve quality of service (QoS) in future wireless networks. A novel measurement-based network selection technique, formulated based on mathematical framework, and terminal-oriented network-assisted (TONA) handover architecture are the main actors of this technology agnostic approach. In particular, QoS parameters estimation is a cornerstone of the generalized CCRRM architecture to facilitate technology abstraction and provide link layer cognition in an effort to realize seamless mobility in future wireless networks. By leveraging on the cooperative exchange of QoS context information over the converged all-IP core and novel concept of reactive QoS balancing (RQB) to achieve the end-to-end goal of promoting a QoS-balanced system, three RQB algorithms augmented with multi-domain cooperation techniques are developed to exploit the heterogeneity of access networks and distribute load opportunistically. Additionally, the radio resource management (RRM) design of the generalized CCRRM architecture is based on a network-terminal distributed decision making process, similar and compliant to the recent IEEE 1900.4 standard. Performance evaluation is conducted with comprehensive discrete event based simulation studies to gain insights of the promising intrinsic benefits associated with RQB under realistic, pragmatic scenarios. Furthermore, an elegant unified analytical model is developed to obtain the key performance metrics for the IEEE 802.11 distributed coordination function (DCF) infrastructure basic service set (BSS), under non-homogeneous conditions, by integrating a Markov chain model in conjunction with a finite queueing model. These performance metrics serve as bounds for reliable capacity analysis from which a model-based predictive QoS balancing (PQB) algorithm is developed as a benchmark for comparative performance studies with the proposed measurement-based RQB algorithm. The contributions of this thesis are not restricted to multiple access point (multi-AP) wireless local area network (WLAN), and the proof of concept is validated based on a heterogeneous multi-AP WLAN where appropriate. Moreover, conditions under which the generalized CCRRM architecture provides abstraction from underlying technologies and stays relevant to future IP-based multi-RAT environment have been established.