Title :
Finite sample identifiability of multiple constant modulus sources
Author :
Leshem, Amir ; Petrochilos, Nicolas ; Van der Veen, Alle-Jan
Author_Institution :
Dept. ITS, Delft Univ. of Technol., Netherlands
Abstract :
We prove that mixtures of continuous constant modulus sources can be identified with probability 1 with a finite number of samples (under noise-free conditions). This strengthens earlier results which only considered an infinite number of samples. The proof is based on the linearization technique of the Analytical Constant Modulus Algorithm, together with a simple inductive argument. We then study the finite alphabet case. In this case we provide an upper bound on the probability of non-identifiability for finite sample of sources. We show that under practical assumptions, this upper bound is tighter than the currently known bound.
Keywords :
array signal processing; blind source separation; identification; linearisation techniques; probability; signal sampling; PSK; analytical constant modulus algorithm; blind source separation; constant modulus signals; continuous constant modulus sources; finite alphabet; finite sample analysis; finite sample identifiability; linearization technique; multiple constant modulus sources; noise-free conditions; nonidentifiability probability; sensors array; upper bound; Algorithm design and analysis; Binary phase shift keying; Constellation diagram; Cost function; Linearization techniques; Sensor arrays; Signal processing; Signal processing algorithms; Upper bound; Vectors;
Conference_Titel :
Sensor Array and Multichannel Signal Processing Workshop Proceedings, 2002
Print_ISBN :
0-7803-7551-3
DOI :
10.1109/SAM.2002.1191071