Estimating the number of received signals by the "deterministic-MDL-criterion"

Estimating the number of received signals by the "deterministic-MDL-criterion"

In this paper we consider estimation methods, which are based on the theory of "minimum description length" (MDL). Considering the proof of the MDL criterion given by G. Schwarz, the "deterministic-MDL-criterion" is derived and compared to the "eigen-MDL-criterion", pro...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
1. Verfasser: Worms, J.G.
Format: Tagungsbericht
Sprache:eng
Schlagworte:
Covariance matrix
Eigenvalues and eigenfunctions
Linear antenna arrays
Maximum likelihood estimation
Probability density function
Sensor arrays
Signal processing
Signal resolution
Stochastic processes
Testing
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:In this paper we consider estimation methods, which are based on the theory of "minimum description length" (MDL). Considering the proof of the MDL criterion given by G. Schwarz, the "deterministic-MDL-criterion" is derived and compared to the "eigen-MDL-criterion", proposed by Wax et. al. Having only a restricted number of sampled data in most cases the new criterion leads to better estimates than the "eigen-MDL-criterion". Our theoretical results are validated by numerical simulations.
ISSN:1097-5659
2375-5318
DOI:10.1109/RADAR.2005.1435828