Distributed sequential detection based on cell-average CFAR local tests
W. Aoya, Z. Shenghua, P. Xiaojun, X. Xiaoyang · IET conference proceedings. · 2023
In conventional distributed radar target detection, the number of accumulation returns from local sensors is typically fixed and then a decision may be made either when the information accumulated is insufficient or when it is enough excessively. In this paper, we generalize the sequential probability ratio test (SPRT) into distributed detection with local sensors using the cell-average constant false alarm rate (CA-CFAR) tests under unknown noise level. The asymptotic approximated thresholds are calculated in the proposed test. Numerical examples show that the exact error probabilities are closer to the given error probabilities, and the exact expected sample sizes (ESSs) are smaller when using the calculated thresholds than those in Wald's thresholds setting. Performance functions of the proposed method are evaluated through Fredholm integral and numerical results show its superiority compared with Wald's SPRT in performance.