Detection of the number of superimposed signals using modified MDL criterion: A random matrix approach
Ahmad Bazzi, Dirk T. M. Slock, Lisa Meilhac · 2016
The problem of estimating the number of superimposed signals using noisy observations from N antennas is addressed. In particular, we are interested in the case where a low number of snapshots L = O(N) is available. We focus on the Minimum Description Length (MDL) estimator, which is revised herein. Furthermore, we propose a modified MDL estimator, with the help of random matrix tools, which improves the estimation of the number of sources. Simulation results demonstrate the potential of the modified MDL estimator over the traditional one, in the case where L = O(N).