Physical layer algorithms with limited feedback: Power control and coding strategies.

摘要

The uncertainty in the received signal is the main limiting factor for the throughput of any communication system with limited resources. In this thesis we analyze the potential gain in throughput by managing channel uncertainty, noise, and interference using algorithms such as power control and coding techniques applied at the transmitter. Our contributions are (i) design and analysis of algorithms based on limited information about the uncertainties, and (ii) taking into account the resources required for collecting these information.; The thesis is divided into two parts. The first part considers the case of channel information at the transmitter. Since most receivers implicitly or explicitly estimate the channel, often a closed loop system is designed to provide the channel state information at the transmitter. We model a finite capacity feedback link by assuming that only few bits of information about the channel are available at the transmitter. We develop a power control algorithm that is asymptotically optimum. Using the designed power control scheme, we characterize the tradeoff between diversity order and multiplexing gain of the multiple antenna systems, and conclude that feedback can significantly reduce uncertainty due to channel fading.; In the second part we consider a multiple access channel with two transmitters and a single receiver, and analyze the impact of cooperation between the senders on the system throughput. In two case studies we establish the importance of cooperation, and the tradeoff between throughput and access overhead. Further, we show that the cost associated with establishing cooperation may exceed the consequent gain. Thus, as the major contribution of the second part, we develop a model based on two-way channel. We are the first to find upper and lower bounds on achievable rates of multiple access two-way channels with noisy feedback. Since the feedback link uses the same channel resources, the cost of feedback is automatically considered. By constructing a coding scheme we show that the two-way channel model is essential in having throughput gain in multiple access channel with noisy feedback. We also demonstrate that feedback based cooperation can reduce uncertainty due to interference and additive noise.