Performance analysis of single receiver Matched-Mode processing for source localization

Yann Le Gall (1) , François-Xavier Socheleau (1) , Julien Bonnel (1)

1 Lab-STICC_ENSTAB_CID_TOMS

Matched-Mode processing (MMP) is an alternative to Matched-Field processing (MFP) that explicitly uses the modal description of the acoustic field to estimate source position and/or oceanic environmental parameters. MMP involves two steps, first the modes are separated, and then they are compared to mode replicas in order to infer the desired parameter values. If the source emits a short time broadband signal, modal separation can be achieved with a single hydrophone using a time-frequency analysis tool called warping. Within this framework, a method that allows to predict MMP performance is provided: the Ziv-Zakai bound is derived, which gives a bound on the achievable performance that is much tighter than the Cramer-Rao bound. The results are then used to perform quantitative performance analyses of the single receiver MMP source localization problem. Particularly, performance with respect to noise is examined for incoherent MMP.

Mots clés: Matched-Mode processing; maximum-likelihood; performance analysis; Ziv- Zakai bound

Domaines: Acoustique [physics.class-ph]; Traitement du signal et de l'image [eess.SP]

Yann Le Gall, François-Xavier Socheleau, Julien Bonnel. Performance analysis of single receiver Matched-Mode processing for source localization. UA 2014, Jun 2014, Rhodes, Greece. ⟨hal-01139856⟩