Random matrices in communications theory - Part I
Tuesday, September 27th, 2005Date: Tuesday, October 4, 2005
Who: Matthew Peacock
Seminar type: Tutorial
Where: Conference Room, Physics Annexe
When: 4-5pm
Abstract:
Modern communications systems are naturally described by linear matrix-vector equations. Essential properties of the system, such as the maximum information throughput and the signal-to-interference ratio of common receivers, are determined by the eigenvalues of the matrices involved. However, the matrices involved are random, which occurs, for example, when the signal from a mobile bounces off trees, buildings, etc., before reaching the base station. Therefore the eigenvalues of the system are also random. Determining the performance of the system is therefore is a complex problem. However, a remarkable thing occurs if you make every parameter (e.g., number of users, number of antennas, etc.) in the system infinitely large, while keeping the ratios of the parameters in proportion: the performance of this asymptotically large system becomes deterministic. Not only that, but the performance of non-infinite sized systems (which are complex to evaluate) is very well approximated by the performance of the infinite sized system.
In these tutorials, I hope to give a brief overview of the main results of random matrix theory which are applicable to communications systems, and show the derivation of a few simple results. I will also give an introduction to free probability theory, and demonstrate some applications of the theory.