Resolution Doubled Co-Prime Spectral Analyzers for Removing Spurious Peaks
Xiangdong Huang, Yiwen Han, Ziyang Yan, Hongyu Xian, Wei Lu · IEEE Transactions on Signal Processing · 2016
Co-prime DFT filter banks provide an effective tool of spectral analysis in case that the sampling rate is far lower than the Nyquist sampling rate. However, since the practical prototype filters are not ideal, spurious peaks will occur in multiple locations when the frequency offsets of the stimulus components are close to half of the frequency resolution of the co-prime spectrum. As a result, the quality of the original co-prime panorama spectrum will degrade greatly. This paper proposes a modified analyzer by synthesizing the original co-prime analyzer and a 0.5 frequency resolution shifted analyzer, which is derived from the original one by applying two simple operations on each low-rate data channel. Following some decision criterion between the original co-prime analyzer and the shifted analyzer, the synthesized analyzer not only removes all the annoying spurious peaks but also doubles the spectral resolution. Numerical comparisons are presented throughout to illustrate the above improvements.