MEVA-Data-Repo GPS

Synopsis

During the KF1 collection, actors and personnel carried GPS loggers set to log their location once every 10 seconds. This directory contains the GPS data corresponding to the time windows of the released MEVA video data. This data is released under the MEVA license.

Consistency of logger IDs

The KF1 collection spans two sessions, one in March and one in May (as of this writing, only data from the March session has been released.) Within each session, each person used the same GPS logger for all days of the collect, with instructions to turn them on at the start of the day's activities and off as they were leaving, resulting in consistent GPS IDs across days within the session. GPS IDs are not consistent between the March and May sessions. Logger IDs are strings of the form Gnnn (e.g. G603.)

GPX file contents

Each of the 116 GPX files represents a five-minute "slot" of wall-clock time, containing all logger tracks during that slot. For example, the file 2018-03-12.10-15-00.gpx contains all tracks on March 12, 2018 from 10:15am to 10:20am.

Relating GPX files to video clips

Each GPX file's time slot (e.g. 2018-03-12.10-15-00) corresponds to entries in the timeslot column in the file meva-clip-camera-and-time-table.txt, described here; this should assist aligning video clips and GPS logs.

Timestamps and time zones

The timestamps in the GPX files are in UTC; all other MEVA timestamps (for example, in filenames) are in MUTC local time (UTC-4).

Filtering of track points

Track points have been cropped to a box around the KF1 collection zone from (39.052214N, 85.534343W) to (39.046152N, 85.524140W); this filters out grossly spurious readings from the logger as well as occasions when an actor may have left the collection zone (e.g. lunch breaks.)

Number of logger IDs per file

The number of unique loggers per file is graphed below. Very low counts correspond to early-in-the-day time slots, as actors arrived on site and turned on their loggers. Variations during any given day have not been thoroughly analyzed but may be due to actors outside the filtering zone, failure to log due to lack of signal (i.e. inside a building), or human error in forgetting to turn the logger on.

POC

Please send questions and comments to mevadata (at) kitware.com.