A compressed sensing sparse channel estimation method for unique word prefixed single carrier frequency domain equalization

In this paper, a compressed sensing (CS) based sparse channel estimation method is proposed for Unique Word (UW) Prefixed SC-FDE employed in sparse wireless channels, sparse channel estimation is formulated as a typical CS problem, and UW generation schemes are discussed, in addition, DantzigD selector is used to recover sparse channel impulse response (CIR) from limited number of noisy observation measurements. Simulation results based on a typical sparse underwater acoustic channel profile show that CS based channel estimation methods outperform the widely utilized time domain and frequency domain least-square (LS) channel estimation methods in bit error rate (BER) and normalized mean-square error (NMSE). The channel estimation accuracy can be significantly improved compared to time domain LS estimation method.