Abstract:

Two important challenges in detecting sparse signals are how to improve the detection accuracy and reduce the computational complexity. The  local likelihood ratio selection(LRSL) procedure was proposed to detect and identify sparse signal segments in one-dimensional noise data. Different from LRS procedure which directly choose candidate intervals from all intervals, the LRSL procedure only considers neighborhoods of those points whose observed data greater than some threshold. Because of the sparsity of the signals, the proposed procedure can greatly reduce the computational complexity. On the other hand, asymptotic results demonstrate that the LRSL procedure can detect weaker signals. The simulation results indicate that the proposed procedure has high detection accuracy and computational efficiency.

Key words: computational complexity; likelihood ratio selection; local likelihood ratio selection; sparse signal detection; detection accuracy