Joint estimation of source number and DOA using simulated annealing algorithm

Authors:
Junli Liang;Ding Liu
Affiliations:
School of Automation & Information Engineering and the School of Computer Science & Engineering, Xi'an University of Technology, Xi'an, China;School of Automation & Information Engineering, Xi'an University of Technology, Xi'an, China
Venue:
Digital Signal Processing
Year:
2010

Citing 9
Cited 1

The propagator method for source bearing estimation

Signal Processing
Reversible Jump MCMC Simulated Annealing for Neural Networks

UAI '00 Proceedings of the 16th Conference on Uncertainty in Artificial Intelligence
Simulated Annealing Using a Reversible Jump Markov Chain Monte Carlo Algorithm for Fuzzy Clustering

IEEE Transactions on Knowledge and Data Engineering
Robust Full Bayesian Learning for Radial Basis Networks

Neural Computation
A computationally efficient algorithm for joint range--DOA--frequency estimation of near-field sources

Digital Signal Processing
Joint Bayesian model selection and estimation of noisy sinusoidsvia reversible jump MCMC

IEEE Transactions on Signal Processing
Bayesian analysis of generalized frequency-modulated signals

IEEE Transactions on Signal Processing
On the virtual array concept for the fourth-order direction findingproblem

IEEE Transactions on Signal Processing
Applications of cumulants to array processing .I. Apertureextension and array calibration

IEEE Transactions on Signal Processing

Cross-ambiguity function domain multipath channel parameter estimation

Digital Signal Processing

Quantified Score

Hi-index	0.01

Visualization

Abstract

The array signal processing problem has to do with two main issues: 1) the determination of source number; and 2) estimation of the directions-of-arrival (DOA) of these sources. In the classical methods, the determination of source number and the estimation of DOA parameters are executed independently rather than jointly. To jointly estimate source number and DOA, this paper proposes a simulated annealing-based algorithm. The key points of this paper are: i) the prior choices for some unknown parameters are calibrated to compute more efficiently; ii) a penalty term for the parameter number is added to the corresponding objective function Additionally to avoid overfitting; and iii) in order to move to the regions of interest quickly, some cumulants of properly chosen sensor outputs from the uniform linear array (ULA) is computed and their moduluses of Fourier transform (MFT) are mapped trickly to the uniform distribution. Finally simulation results are presented to validate the performance of the proposed method.