Page 17 - Acoustics Today Summer 2011

Fig. 10. Geoacoustic and source tracking example from SWellEx-96 experiment. (a) Towed source trajectory and a vertical line array (VLA) in a range dependent environ- ment with varying geoacoustic parameters and (b) posterior probability density functions (PDFs) of water depth and sediment properties [25] obtained using a particle fil- ter with five minute snapshots provided on the right. Solid curves are the true values calculated from the Bachman database.
Using the same approach, amplitudes of distinct paths can be readily estimated as well, in addition to their arrival times,26 and can be subsequently used for estimation of atten- uation.
Source and geoacoustic tracking
Another application of Model-Based ocean acoustic sig- nal processing is the tracking of source and ocean environ- mental properties using sequential Bayesian techniques. These techniques provide a suitable framework for sequen- tially estimating in time and space the unknown ocean envi- ronment and source parameters as data become available.
KFs deal with systems where the acoustic measurements are related to unknown parameters via linear equations and all the underlying uncertainty can be represented with Gaussian PDFs. Geoacoustic tracking involves nonlinear interactions between the environmental parameters and the measured acoustic field with non-Gaussian PDFs. Hence, EKFs, UKFs, and PFs were used to handle the nonlinear/non-Gaussian
24
geoacoustic tracking problem. The performances of the
EKF, UKF, and PF were analyzed and compared to the Bayesian Cramér-Rao lower bound, that provides a lower bound for achievable performance in tracking geoacoustic
Would you li
ik
ke
e
to meet the n n
e ew
n
?
?
w
head in town
Visit ansihead.com if you are interested in measuring the insertion-loss of all types of hearing protectors.
We make microphones ansihead.com
Model-Based Ocean Signal Processing 13