Channel estimation using compressed sensing via SDR’s

Abstract

Channel estimation is applied at the receiving end to get the reaction of channel so that effects of the channel can be calculated. A noble signal processing technique called compressed sensing is used for channel estimation. We are exploiting the sparsity of the channel in time domain by selecting the pilot symbols or preambles randomly and constructing a random projection measurement matrix. We are using development kit GNU Radio which is software based. We are building a Channel Estimator which is based on Compressed Sensing and implemented on a Universal Software Radio Peripheral 1 (USRP1). Results and plots for compressed channel sensing will be shown to prove the better and effective conclusions as compared to earlier used channel estimation. The approach we are following improves the channel estimation accuracy by saving the bandwidth effectively. We are developing a practical compressed sensing-based channel estimator. CS can be effective when high quality results/resolution is required and also when we have hardware limitations. Our project demonstrates that how important is the usage of SDRs to fill the gap between a theoretical model and practical implementations, which as a result boosts us to continue following various algorithms.