An improved pilot routing algorithm for compressed sensing-based channel estimation in underwater acoustic OFDM system

Abstract

Compressed sensing (CS) is used to estimate channel information in underwater acoustic OFDM system recently. However, there are no recognition criteria to specify the pilot patterns in channel estimation. In order to improve the accuracy of the estimation, this paper proposes a pilot routing algorithm based on minimizing the adjusted mutual coherence of the measurement matrix in the recovery algorithm of compressed sensing theory. Each pilot symbol location is determined by calculating the adjusted mutual coherence of the corresponding measurement matrix. The matrix is produced by fixing other pilot location indexes and replacing the index of the current pilot location with all optional indexes. The simulation results show that the BER is 2dB and 8dB lower when SNR is 5dB and 30dB respectively, and the MSE is 2dB lower compared with the results of algorithm 1. Comparing with the results of algorithm 2, the BER is 2dB better, and the MSE is 3dB lower when SNR is 30dB. The data confirms the efficiency of this pilot routing algorithm.