Premium
Performance of a space‐based wavelet compressor for plasma count data on the MMS Fast Plasma Investigation
Author(s) -
Barrie A. C.,
Smith S. E.,
Dorelli J. C.,
Gershman D. J.,
Yeh P.,
Schiff C.,
Avanov L. A.
Publication year - 2017
Publication title -
journal of geophysical research: space physics
Language(s) - English
Resource type - Journals
eISSN - 2169-9402
pISSN - 2169-9380
DOI - 10.1002/2016ja022645
Subject(s) - lossless compression , data compression , wavelet , compression (physics) , data compression ratio , computer science , compression ratio , plasma , spectrometer , optics , algorithm , physics , image compression , artificial intelligence , image processing , quantum mechanics , internal combustion engine , image (mathematics) , thermodynamics
Abstract Data compression has been a staple of imaging instruments for years. Recently, plasma measurements have utilized compression with relatively low compression ratios. The Fast Plasma Investigation (FPI) on board the Magnetospheric Multiscale (MMS) mission generates data roughly 100 times faster than previous plasma instruments, requiring a higher compression ratio to fit within the telemetry allocation. This study investigates the performance of a space‐based compression standard employing a Discrete Wavelet Transform and a Bit Plane Encoder (DWT/BPE) in compressing FPI plasma count data. Data from the first 6 months of FPI operation are analyzed to explore the error modes evident in the data and how to adapt to them. While approximately half of the Dual Electron Spectrometer (DES) maps had some level of loss, it was found that there is little effect on the plasma moments and that errors present in individual sky maps are typically minor. The majority of Dual Ion Spectrometer burst sky maps compressed in a lossless fashion, with no error introduced during compression. Because of induced compression error, the size limit for DES burst images has been increased for Phase 1B. Additionally, it was found that the floating point compression mode yielded better results when images have significant compression error, leading to floating point mode being used for the fast survey mode of operation for Phase 1B. Despite the suggested tweaks, it was found that wavelet‐based compression, and a DWT/BPE algorithm in particular, is highly suitable to data compression for plasma measurement instruments and can be recommended for future missions.