A Modified Time Averaging Method in Weak Signal Detection
Xiang Jin-lin · Tance yu kongzhi xuebao · 2003
This paper deals with a modified time averaging method for the detection of weak periodical signal. When the struncating period deviates the integral times of the signal period, the processing gain of time averaging deteriorates rapidly. However, noise added on the signal is attenuated at the same time. Our analysis indicates that noise attenuates faster than the signal does before the averaging times reach a threshold. If we manage to obtain the optimal averaging times, we can still get fairly high processing gain. Based on the above analysis, we propose a method to search weak periodical signal in a limited frequency band by changing the struncating period. Simulation shows that this method can detect weak periodical signal under a very low SNR of -35dB.