Detection of Signal with Unknown Parameter in Colored Noise
Hongmei Xie · Jisuanji fangzhen · 2009
Time series are as a sum of a signal with unknown parameter and a wide sense stationary Gaussian noise is studied,and autoregressive(AR) model is used for the noise.Least square estimation of the AR and signal parameters are a nonlinear estimation problem.With the property of subspace invariance of the signal,it can be reduced to a linear LS problem.Using the matched subspace method,the paper obtained the generalized likelihood ration detector(GLRT) and the probability of detection and false alarm.Theoretical analysis and computer simulation proved the effectiveness of the method.