Cooperation and Cognition offer new exciting dimensions to future wireless technology. In a broad sense, cooperative networks refer to those systems where transmitting nodes share their resources, such as antennas, to improve coverage, diversity and spectral efficiency. Cognition, on the other hand, refers to discovery of unused resources, such as spatially or temporally unused spectrum holes, to allow regulated coexistence of low and high priority users. Moving from a rigid regulatory framework of competitive/exclusive usage of resources to a cooperative/shared resource-pool paradigm is challenging due to numerous uncertainties arising from practical system considerations.;In this thesis, we focus on the analysis and modeling of such uncertainties, as well as, the development of robust, yet efficient strategies, given informational constraints in the context of cooperative relay networks and spectrum agile cognitive radio systems. In particular, we consider Geometry Induced Uncertainties (ambiguity in transmission and observation channel models), Side Information Uncertainties (imperfections in channel knowledge at the transmit nodes) and Behavioral Ambiguity of coexisting users. Addressing these sources of uncertainty involved: (1) Proving the existence of non-parametric transmission schemes in cooperative relay networks, i.e., schemes which are universally optimal for a wide range of smooth channel probability distribution functions; (2) Development of robust strategies that only make use of statistical information about the channel in multi-user multi-relay setups; (3) Developing cooperative sensing strategies that adopt new statistical performance metrics to deal with lack of full observation models in spectrum sensory networks; (4) Use of hierarchical coding to cutoff spectrum violators while reducing rates of false unsubstantiated allegations and accusations of disloyalty. Our analytical/experimental evaluations suggest that: (i) Even with imperfect CSI, using our strategies results in significant throughput improvements when compared to conventional non-cooperative schemes with achievable throughput that is fairly close to the optimum (genie-aided) throughput; (ii) The twin objectives of bandwidth utilization and interference control can significantly benefit from group testing across all channels in contrast: to conventionally employed channel-by-channel detection strategies while reducing the sensitivity requirements of cognitive radios. Our results allow to investigate the impact of the quality of feedback on the system throughput of cooperative relay networks, and thus trade off overhead for the feedback channel with throughput during the actual data transmission. They also provide solutions for spectrum agile networks towards trustworthy coexistence environments with legacy users protection and better secondary performance tradeoffs.
展开▼
