UC San Diego
Feedback, power control, and beamforming : methods for situational aware wireless networks
- Author(s): Huang, Yichao
- et al.
Feedback, power control and beamforming are important PHY and MAC layer issues in wireless communication systems. This dissertation is concerned with theoretical studies and algorithmic developments related to these issues in the context of and in support of the design of situational aware wireless networks. The envisioned situational aware wireless networks adapt system parameters and algorithms design to the channel attributes, user attributes, and system attributes, which constitute the wireless environment and network situations. The research topics in this dissertation regarding feedback, power control and beamforming are motivated by issues that arise from considering different types of awareness. We first propose an adaptive feedback design based on the heterogeneous spectral channel statistics among users, which advocates the awareness of channel attributes. We leverage the multi -cluster subband fading model to develop an analytical framework to investigate the impact of partial feedback and potential imperfections including channel estimation error and feedback delay on system performance. Next, we examine partial feedback in a heterogeneous multicell, and propose a heterogeneous feedback design based on heterogeneous user densities and large scale channel effects, which advocates the awareness of user attributes. The cumulative distribution function (CDF)-based scheduling policy is employed to obtain multiuser diversity gain while maintaining scheduling fairness. We derive a closed form expression as well as asymptotic approximation for the sum rate. In addition, the CDF-based scheduling policy is leveraged in a random beamforming framework to address several open problems. We develop the notion of individual sum rate to study the rate scaling for each individual user. We theoretically examine the randomness of multiuser diversity incurred by selective feedback to further establish the individual scaling laws under different feedback schemes. We then investigate joint beamforming and power control in a multiuser interference network, and propose the usage of spatial channel statistics for algorithm design, which advocates the awareness of channel attributes. With the outage event induced by the utilization of spatial channel statistics, we present decentralized and fast convergent algorithms to achieve outage balancing in the interference network. Finally, we study joint beamforming and power control in a coordinated multicell downlink and employ the max-min formulation to enforce egalitarian user fairness. In order to design efficient algorithm that scales well with the system dimension, we leverage the large system structure and advocate the awareness of system attributes. In our proposed algorithm design, the asymptotic power is computed using statistical channel information and the instantaneous beamformer is obtained in a non-iterative manner. We also establish the effective network to characterize and interpret the asymptotic solution