[Invited Talk] Applications of Compressed Sensing in Wireless Communication Systems

摘要

Compressed sensing has been drawn explosive attention during last several years. It is a new framework that uses signal sparsity to reduce the amount of data that needs to be measured. The measurement of an M-dimensional vector onto an N-dimensional (N＜＜M) vector loses some information in general and the inverse problem (N equations with M variables) has an infinite number of solutions. However, when the signal only has K nonzero coefficients (K＜N＜＜M) in some convenient basis and the measurement matrix is incoherent with the basis, or equivalently satisfies the restricted isometry property, the inverse problem has, with high probability, a unique and exact solution. Inspired by above seminal work of Candes and Donoho, and a plethora of related research has taken place from theory to applications including astronomy, magnetic resonance imaging, and digital imaging by exploiting the natural sparsity of their underlying data. Interpretation of compressed sensing in the sense of wireless technology is as follows: radio wave data can be received, transmitted, and reconstructed using sub-Nyquist rate information without aliasing if the original signal is sparse in time domain or frequency domain.