Application of wavelet package transform in sweep signal active detection

In this paper a TFCC (time-frequency cross correlation) algorithm is put forward which can be used in sweep signal detection and its time delay estimation in distance measurement and active detection. The source sweep signal and received signal of active detection are transformed into time-frequency domain with wavelet package transform, and time-frequency cross correlation between the source sweep signal and the received signal in time-frequency domain is performed. The result is a series of time-frequency correlation wavelet which can indicate the existence of the reflections and their time delay. In order to improve the performance of reflection detection, non-downsample Mallat algorithm was used in decomposition of the signals in wavelet package transform. The TFCC algorithm can suppress noise very effectively because it can avoid the interference of the ambient noise to the reflections and the interference between the reflections to each other. The results of numerical experiment shows that the TFCC algorithm can estimate the time delay of sweep signal accurately up to SNR of -50 dB, while the cross correlation can only do up to SNR of -12 dB.