Barankin lower bound for change-points in independent sequences

In this paper, the computation of the Barankin bound (BRB) for an unbiased estimator of the step location for an independent discrete-time signal is proposed. In the derivation of the change-point BRB for different observation vector distributions the usual Gaussian case is first considered. Simulation results show that the BRB provides a reasonable lower bound for the estimated mean square errors (MSE) of the change-point maximum likelihood estimator (MLE).

[1]  J. Hammersley On Estimating Restricted Parameters , 1950 .

[2]  Alfred O. Hero,et al.  On Achievable Accuracy in Edge Localization , 1992, IEEE Trans. Pattern Anal. Mach. Intell..

[3]  A. Swami,et al.  Step-change localization in additive and multiplicative noise via multiscale products , 1998, Conference Record of Thirty-Second Asilomar Conference on Signals, Systems and Computers (Cat. No.98CH36284).

[4]  H. Messer,et al.  Analysis of inherent limitations in localizing step-like singularities in a continuous signal , 1996, Proceedings of Third International Symposium on Time-Frequency and Time-Scale Analysis (TFTS-96).

[5]  Ali M. Reza,et al.  Cramer-Rao lower bound on locations of sudden changes in a steplike signal , 1996, IEEE Trans. Signal Process..

[6]  Zhen Wang,et al.  All-purpose and plug-in power-law detectors for transient signals , 2001, IEEE Trans. Signal Process..

[7]  Brian M. Sadler,et al.  Cramer-Rao bounds for step-change localization in additive and multiplicative noise , 1998, Ninth IEEE Signal Processing Workshop on Statistical Signal and Array Processing (Cat. No.98TH8381).

[8]  Brian M. Sadler,et al.  Analysis of Multiscale Products for Step Detection and Estimation , 1999, IEEE Trans. Inf. Theory.