Adaptive algorithm for independent component analysis of complex valued signals
Yanfei Jia · Harbin Gongcheng Daxue Xuebao/Journal of Harbin Engineering University · 2009
To overcome the strong uncorrelated transform algorithm's inapplicability to noncircular signals with the same spectral coefficients,an adaptive algorithm with independent component analysis(ICA) for complex valued signals was proposed.A cost function was constructed for this algorithm based on the property that second order statistics of noncircular signals were not zero.By optimizing the cost function directly with a complex valued gradient,an adaptive algorithm for noncircular signals was derived.This algorithm is not only effective for noncircular signals with the same spectral coefficients,but is also applicable to any statistically independent source signal,including noncircular signals,without requiring computation of eigenvectors and singular value vectors.Simulations proved the effectiveness of this new algorithm.