Blind Constant False Alarm Detection Algorithm for Broadband Multiple Signals

Aiming at the problem of broadband multi-target signal detection, an iterative blind detection algorithm based on sparse Bayesian is proposed. Under the condition of unknown signal number and spectrum position, the proposed method can realize blind detection of multi-target signal with constant false alarm (CFAR) probability. Based on the analysis of signal characteristics, a linear model is established by constructing an overcomplete dictionary matrix, and the multi-signal detection is transformed into a non-zero row discrimination problem of the channel coefficient matrix for processing. Based on the sparse characteristic of signal frequency domain distribution, the joint solution is realized by iteration under the sparse Bayesian framework. The analytical expression of the detection threshold is derived theoretically. Based on the parameter estimation results, the multi-target signal CFAR blind detection is realized through automatic zero pruning. Simulation results show that the proposed algorithm has strong flexibility and robustness, and has obvious performance advantages over existing sparse detection methods under the condition of short data lengths.