A CFAR based model order selection criterion for complex sinusoids

Authors:
Changzheng Ma;Boon Poh Ng
Affiliations:
School of Electronic Information, Zhongyuan University of Technology, Zhengzhou, Henan, China;School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore
Venue:
Signal Processing - Signal processing in UWB communications
Year:
2006

Citing 3
Cited 1

A model selection rule for sinusoids in white Gaussian noise

IEEE Transactions on Signal Processing
A combined order selection and parameter estimation algorithm forundamped exponentials

IEEE Transactions on Signal Processing
Asymptotic MAP criteria for model selection

IEEE Transactions on Signal Processing

Adaptive subspace detection of range-distributed target in compound-Gaussian clutter

Digital Signal Processing

Quantified Score

Hi-index	0.00

Visualization

Abstract

The model order selection problem for complex sinusoidal signals has often been tackled using Akaike information criterion (AIC), minimum description length principle (MDL) or maximum a posterior criterion (MAP). As a finite-length noise sequence (discrete-time noise signal) can be expressed as a summation of sinusoidal signals, we propose signal-energy-to-noise-power-ratio (SENR) to be considered in the model order selection problem. In our proposal, a predefined SENR threshold is used to decide whether a signal exists or not and the signal order can then be obtained. Since this new criterion is just like the constant false alarm rate (CFAR) detector, we call the new criterion the CFAR criterion. In this paper, the performances of the AIC, MDL, MAP and CFAR criteria are compared. Simulation results testify the correctness of our analysis and the superior performance of the CFAR criterion.