Low Complexity Sparse Channel Estimation Based on Compressed Sensing

Abstract

In wireless communication, channel estimation is a key technology to receive signal precisely. Recently, a new method named compressed sensing (CS) has been proposed to estimate sparse channel, which improves spectrum efficiency greatly. However, it is difficult to realize it due to its high computational complexity. Although the proposed Orthogonal Matching Pursuit (OMP) can reduce the complexity of CS, the efficiency of OMP is still low because only one index is identified per iteration. Therefore, to solve this problem, more efficient schemes are proposed. At first, we apply Generalized Orthogonal Matching Pursuit (GOMP) to channel estimation, which lower computational complexity by selecting multiple indices in each iteration. Then a more effective scheme that selects indices by least squares (LS) method is proposed to significantly reduce the computational complexity, which is a modified method of GOMP. Simulation results and theoretical analysis show the effectivity of the proposed algorithms.