Chaumette, Eric and Galy, Jerome and Vincent, François and Larzabal, Pascal Computable lower bounds for deterministic parameter estimation. (2007) In: Computational Advances in Multi-Sensor Adaptive Processing, 2007. CAMPSAP 2007, 12-14 Dec 2007, Virgin Islands, United States.
|(Document in English) |
PDF (Author's version) - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Official URL: http://dx.doi.org/10.1109/CAMSAP.2007.4497960
This paper is primarily tutorial in nature and presents a simple approach(norm minimization under linear constraints) for deriving computable lower bounds on the MSE of deterministic parameter estimators with a clear interpretation of the bounds. We also address the issue of lower bounds tightness in comparison with the MSE of ML estimators and their ability to predict the SNR threshold region. Last, as many practical estimation problems must be regarded as joint detection-estimation problems, we remind that the estimation performance must be conditional on detection performance, leading to the open problem of the fundamental limits of the joint detectionestimation performance.
Repository Staff Only: item control page