Performance of Bayes-Optimal Angle-of-Arrival Estimators
Abstract
The angle-of-arrival estimation problem for waves incident upon a sensor array was examined through a Monte Carlo evaluation of the performance of the Bayes-optimal MAP (maximum aposteriori) and MMSE (minimum mean square error) estimators. The case of two independent wave emitters of known powers as well as a multiple look, Gaussian signal in Gaussian noise statistical model were assumed. The Cramer-Rao bound on the estimator's rms error was computed for comparison. The evaluation proceeded with the computation of MAP and MMSE angle estimates for 1000 random samples of array outputs and the accumulation of their rms errors. The probability of detecting both emitters with the optimal detector was also accumulated. This was done for .1, .03, and .01 beamwidths emitter separations and a range of signal-to-noise ratios (SNRs). The accuracy of the computations was assured through a simple finite grid approximation for the estimates, with no covergence problems, and through the evaluation of statistical confidence intervals for the Monte Carlo data. Additional results included properties of the aposteriori probability density and an analytical computation of the performance of the known angles-of-arrival optimal detector.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 13, 1984
- Accession Number
- ADA146594
Entities
People
- F. M. White
Organizations
- Massachusetts Institute of Technology