Abstract: A new time delay estimation algorithm for weak narrowband signals based on sparse reconstructive was presented. Firstly, the data matrix was constructed by the cross-correlation spectrum of signals, secondly, the over-complete dictionary for time delay was established, finally, the high precision time delay was estimated in the signal subspace by using orthogonal matching pursuit (OMP) algorithm, and the subspace is obtained by singular value decomposition (SVD) of the data matrix. Analysis and simulation verify the correctness and effectiveness of the algorithm. Compared with conventional time delay estimation algorithm, the presented algorithm can improve the estimation accuracy by about 1 time.

Key words: time delay estimation, sparse reconstructive, orthogonal matching pursuit, weak narrowband signals