An Eigenvalue Decomposition-Based Method for In-Service Testing of Wireless Communications Systems

Reliably measuring the relevant parameters (like channel power, occupied bandwidth, and carrier frequency) characterizing the signals involved in wireless communications systems turns out to be the key for their in-service testing at a physical layer. This task, however, would not succeed whenever other signals, generated by different wireless systems, interfere with the useful one both in the time and frequency domains. To overcome this limitation, the authors have recently assessed the suitability of eigenvalue decomposition-based algorithms in accurately estimating the power spectrum of a given number of signals that overlap in the time domain and occupy, fully or partially, the same frequency band. Starting from the promising outcomes of the cited algorithms and taking advantage of some nice features of information theoretical criteria, a new measurement method is designed and implemented hereinafter. Designed directly for the in-service physical layer testing of wireless communications systems, the method allows a blind separation of the power spectrum of the useful signal from those of the interfering ones. The evaluation of the desired parameters of the useful signal as well as the frequency-domain characterization of the interfering ones is thus possible, whatever the number of involved signals. The results obtained in a number of laboratory experiments under different in-channel interference conditions show the reliability and efficacy of the method also in the presence of severe in-channel interference.