Sparse Bayesian learning for a bilinear calibration model and mismatched CRB

Gopala, Kalyana; Kurisummoottil Thomas, Christo; Slock, Dirk TM
EUSIPCO 2019, 27th European Signal Processing Conference, 2-6 September 2019, Coruna, Spain

Variational Bayesian (VB) estimation allows for approximate Bayesian inference. It determines the closest approximation in factored form of the posterior distribution by
minimizing the Kullback-Leibler distance to the posterior distribution even if this last one is difficult to determine. In spite of this well motivated derivation, the performance of VB techniques is not very clear, especially compared to more classical performance
bounds. In this paper we explore recently introduced mismatched Cramer-Rao bounds (mCRB) for Bayesian estimation in the context of VB estimation. We focus on the case of bilinear signal models. One particular application of these models arises in the
context of internal relative reciprocity calibration of Massive antenna arrays, in which the received signals are linear in terms of an intra array channel and the relative calibration factors. We have recently shown that a VB approach allows for particularly improved estimation performance that goes beyond the classical CRB, which is now confirmed by the mCRB.

Systèmes de Communication
Eurecom Ref:
© 2019 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.