Compressive Joint Angular-Frequency Power Spectrum Estimation for Correlated Sources  

Oleh:

Dyonisius Dony Ariananda ; Leus, Geert

Jenis: Article from Proceeding
Dalam koleksi: Prosiding Seminar Sistem Telekomunikasi dan Informasi (SSTI) 2014 Jakarta, 27 - 28 Oktober 2014, page 5-10.
Topik: Multi-coset sampling; dynamic linear array; power spectrum; wide-sense stationary
Fulltext: Prosiding_SSTI_2014-2.pdf (1.14MB)

Isi artikel

In this paper, the power spectrum of the receivedwide-sense stationary (WSS) signals produced by possibly correlatedsources is compressively estimated as a function of directionof arrival (DOA) and frequency. Here, we compress the receivedWSS signals in both the time and the spatial-domain. Thespatial-domain compression is performed by using the so-calleddynamic linear array of active antennas and the time-domaincompression is implemented by sampling the signal receivedby each active antenna at sub-Nyquist rate using multi-cosetsampling. In each time slot, we then calculate the correlationsbetween the resulting sub-Nyquist rate samples both in thespatial-domain and the time-domain and use them to recoverthe two-dimensional (2D) power spectrum matrix that gives thepower spectrum information in both the frequency and theangular dimensions. Under the full rank condition of the systemmatrices, the 2D power spectrum reconstruction can generally bedone using simple least squares approach without applying anysparsity constraint on the true power spectrum. In addition, theresulting angular power spectrum at each frequency can also beused for DOA estimation of more correlated sources than activesensors by locating the peaks in the power spectrum.

Opini Anda

Klik untuk menuliskan opini Anda tentang koleksi ini!

Kembali