Recursive hybrid CRB for Markovian systems with time-variant measurement parameters

Jerome Galy, Alexandre Renaux, Eric Chaumette, Francois Vincent, Pascal Larzabal
2015 2015 IEEE 6th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing (CAMSAP)  
In statistical signal processing, hybrid parameter estimation refers to the case where the parameters vector to estimate contains both deterministic and random parameters. Lately computationally tractable hybrid Cramér-Rao lower bounds for discrete-time Markovian dynamic systems depending on unknown time invariant deterministic parameters has been released. However in many applications (radar, sonar, telecoms, ...) the unknown deterministic parameters of the measurement model are time variant
more » ... ich prevents from using the aforementioned bounds. It is therefore the aim of this communication to tackle this issue by introducing new computationally tractable hybrid Cramér-Rao lower bounds. Therefore, using block matrix inversion: Moreover, for MDS, (8) leads to: p (y 1:k , x 0:k |θ k ) = p (y k |x k , λ k , µ k , θ 0 ) p (x k |x k−1 , θ 0 ) × p y 1:k−1 , x 0:k−1 |θ k−1 yielding:
doi:10.1109/camsap.2015.7383839 dblp:conf/camsap/GalyRCVL15 fatcat:b5lnnniqwjchvaztdx6d3h2blu