headgraphic
loader graphic

Loading content ...

Compression and Assimilation for Resource-Limited Operations

P.F.J. Lermusiaux, P.J. Haley, Jr.,
T. Ryu, C. Mirabito

Massachusetts Institute of Technology
Center for Ocean Engineering
Mechanical Engineering
Cambridge, Massachusetts

Co-PIs:
D. Walker
SRI International

J. Fabre, G. Jacobs, J. Metzger
NRL Stennis Space Center
Project Summary
Ongoing MIT-MSEAS Research
Additional Links
MSEAS Project-supported Publications
Background Information

 

This research is sponsored by the Office of Naval Research.

Project Summary

The goal of the proposed program is to provide environmental situational awareness to forces operating in communications-disadvantaged settings, with a focus on the undersea acoustic environment. The objectives of the proposed program are to mature and demonstrate 1) techniques for compression of ocean forecast fields that maintains critical acoustic properties, 2) tools for assimilation of local observation data to provide improved estimates of present conditions, 3) to conceptualize methods (i.e. analogous environments) for accurately characterizing ocean conditions via identification of matching situations within the reanalysis archive, and 4) to explore approaches using machine learning for reduced-order modeling and forecasting.

Background information is available below.

Top of page

Ongoing MIT-MSEAS Research

Our MIT-MSEAS research will consist of four research thrusts involving reduced order modeling (ROM): i) decomposition of deterministic and probabilistic forecasts for efficient compression, reduction and reconstruction; ii) machine learning for reduced-order modeling and forecasting; iii) adaptive and data-assimilative reduced-order modeling and forecasting; and iv) multi-resolution and multi-dynamics ROMs.

Publications

MSEAS Project-supported Publications

Top of page


Additional Links

Top of page


Background Information

One area of research for this project will encompass data compression approaches such as principal component analysis (PCA) and K-means singular-value decomposition (K-SVD) with a focus on data compression approaches that will enable efficient transmission of ocean forecast and acoustic data to users over bandwidth-limited, disadvantaged communications links. Another focus will be development and assessment of a lightweight PCA-based data assimilation approaches with modest computing requirements to enable on-board use of local data to improve the accuracy and hence utility of environmental forecasts.

A second area of research will encompass methods to provide environmental forecasts with high compression via use of analogous environments. We envision use of a machine learning approach at the Fleet Numerical Meteorology and Oceanography Center (FNMOC) to identify appropriate times in the reanalysis database that are analogous to the present forecast. The information transmitted to a disadvantaged platform is then very small, and the platform can retrieve the analogous environment from the copy of the database onboard. We will also explore machine learning for reduced-order modeling and forecasting.

The third area of work relates to data sources, and assessment of performance for the algorithms being developed. We will use the Global Operational Forecast System (GOFS) 3.1 Reanalysis as a data source for developing compression and assimilation algorithms. Areas of application will include the North Atlantic and the West Pacific, with a focus on characterizing the acoustic environment. The fidelity of the resulting acoustic fields will be assessed using standard approaches developed at NRL, and the assimilation performance will be compared to that obtained using conventional approaches as well as independent observation data.

Top of page


Go to the MSEAS home page