Combine Kart Truck GPS Data Archive

Listed in Datasets

By Yaguang Zhang1, James Krogmeier1

Purdue University

GPS data for agricultural vehicles collected during wheat harvesting.

Additional materials available

Version 1.1 - published on 28 May 2020 doi:10.4231/GMH9-8X88 - cite this Archived on 28 Jun 2020

Licensed under CC0 1.0 Universal

BackTraceabilityTreeOverviewWiden.png All.png TraceTreeResultAllColoredByEle.png Trial5_3_SnippetToDrawDiffFieldShapes_2014_Legend.png


Global Positioning System (GPS) logs during key agricultural operations are crucial for digital agriculture. From them, a massive amount of logistics information could be extracted for aiding farmers in decision making and operation performance improvement. Using an Android app we developed, we collected GPS data from agricultural vehicles during wheat harvesting (2014, 2016, 2017, 2018 and 2019) of a farm located in Colorado, USA.

This dataset includes two zip packages containing the raw GPS data collected by our Android app CKT, together with some post-processing results for them. We have been developing fully-automatic algorithms on high-precision field shape generation, vehicle activity recognition, and product traceability based on these data. For field shape generation, we compared our algorithm’s outputs with corresponding boundaries from the field owner. Resulted field shapes agree really well with these boundaries, but they also capture more details on which exact parts have been harvested. Furthermore, key activities including harvesting and product loading/unloading were recognized, with a satisfactory accuracy compared with manual labels. From these results, a demo product traceability system was implemented to track product all the way from field segments to elevators. In this way, up-to-date high-level knowledge can be automatically extracted from GPS records and help farmers make better logistic decisions.

Source code for the CKT app is publicly available at:

Matlab post processing scripts are publicly available at:

A note about Matlab files: The .mat files were generated over a number of years using different versions of Matlab. The files should be compatible with any Matlab newer than (and including) version 2014b. More specifically, the .mat files holding the GPS data are version 7 .mat files, and they should be compatible with any modern Matlab installations released after 2004 (MATLAB Version 7 or later). However, the field shapes in our post-processing results use an object called “alpha shape”, which was introduced in Matlab 2014b.

Cite this work

Researchers should cite this work as follows:



Update on 05/27/2020: added data for year 2019; updated the readme files accordingly.

Initial publication on 12/19/2018: included data for years 2014, 2016, 2017, and 2018; modified the readme files.

The Purdue University Research Repository (PURR) is a university core research facility provided by the Purdue University Libraries, the Office of the Executive Vice President for Research and Partnerships, and Information Technology at Purdue (ITaP).