ZeroG

Code for working with accelerometer data from parabolic flights.

We use two-stage change-point detection to identify transitions between g-levels and periods of statistically stable g-levels outside these transitions.

Compatibility

Tested with MATLAB 2017a on OS X 10.12 and Windows 7 and Windows 10. Expected to work with MATLAB 2016a+. Requires Signal Processing Toolbox.

Citation

Carr CE, Bryan NC, Saboda KN, Bhattaru SA, Ruvkun G, Zuber MT. Acceleration Profiles and Processing Methods for Parabolic Flight. npj Microgravity 4, Article number: 14 (2018) https://doi.org/10.1038/s41526-018-0050-3. Preprint: arXiv:1712.05737 https://arxiv.org/abs/1712.05737

Installation

Get Scripts

Download: https://github.com/CarrCE/zerog/archive/master.zip or use command line git clone git@github.com:CarrCE/zerog.git.

Unzip to preferred location, here denoted /zerog-master.

Get Data

Download: https://osf.io/5rqu9/download. This 1.0 GB (compressed ZIP) dataset has a CC BY 4.0 US license. More details at: https://osf.io/nk2w4/

Unzip into the same folder as your code, e.g. /zerog-master

You should now have two folders in your /zerog-master directory, Flight and Lab, which contain the flight data and a short lab test used in verifying accelerometer calibration accuracy.

Run Analysis

In MATLAB, go to your /zerog-master path, and run the main script: analysis. This will perform the same analysis as in the publication (see citation, above). For documentation and to adapt to your own data, see the contents of the analysis.m script.

The results of running this analysis in MATLAB include a series of PDF figures, replicating those in the preprint, the filtered g-level data, and a tab-delimited file of all periods in the flight. All times are elapsed time, and for reference, the start time is: 2017-11-17 18:28:51 UTC. The analysis results are also available (188 MB ZIP) at: https://osf.io/pmhj4/download

License

Distributed under an MIT license. See LICENSE for details.