Title :
On blind system identification from its output autocorrelation
Author :
Chen, Yuanjie ; Nikias, Chrysostomos L.
Author_Institution :
Signal & Image Process. Inst., Univ. of Southern California, Los Angeles, CA, USA
Abstract :
When the input to a linear time-invariant system is a sequence of equispaced Dirac delta functions with IID. random amplitudes, it has been shown that a time-limited system can be uniquely identified from its output autocorrelation by making use of the cyclostationarity of the output [Tong et al. 1991]. The present authors show that a bandlimited system, however, can not be uniquely identified from its output autocorrelation. A necessary and sufficient condition for a bandlimited system to be uniquely identified from its output autocorrelation is given. Further, simulation results show that the identification method in Tong et al. does not yield satisfactory results for bandlimited systems
Keywords :
correlation theory; identification; matrix algebra; signal detection; IID random amplitudes; bandlimited system; blind system identification; cyclostationarity; equispaced Dirac delta functions; input; output autocorrelation; time-limited system; Autocorrelation; Bandwidth; Contracts; Ear; Frequency; Image processing; Signal processing; Sufficient conditions; System identification; White noise;
Conference_Titel :
Signals, Systems and Computers, 1993. 1993 Conference Record of The Twenty-Seventh Asilomar Conference on
Conference_Location :
Pacific Grove, CA
Print_ISBN :
0-8186-4120-7
DOI :
10.1109/ACSSC.1993.342617