Emitter number estimation from pulse envelope using information theoretic criterion

In this paper, an approach for estimating the number of emitters from a set of interleaved pulses trains is proposed. The approach is based on the application of information theoretic criterion, which is formulated by using a new model of eigenvalues from principal component analysis (PCA) of pulse envelope vectors. In this model, the logarithm likelihood function is obtained by clustering the eigenvalues into two groups: signal and noise component group. The experimental results suggest that the present likelihood function can provide a good estimate of the dimension of signal component group from artificial data. When compared with the other information theoretic criterions, the proposed information theoretic criterion does not involve any computationally sophisticated maximum likelihood function. In addition, it is simple, intelligible, and more efficient. Computer simulations are used to show the effectiveness and feasibility of the proposed approach.