The Serval Maps application was recently used during KiwiEx 2012 exercise. Dr Paul Gardner-Stephen, project co-founder and lead researcher, attended the exercise. He has a few posts up already about his experiences  and .
This was an exciting opportunity to field test the Serval Maps application and some valuable lessons were learned, including:
- I can respond to feature requests from New Zealand via satellite phone and update the Serval Maps application successfully;
- There is no substitute for field testing with a number of devices, there were a couple of small issues which were highlighted during the testing; and
- In general the application worked well and collected a significant amount of data.
The Serval Maps application uses binary files and the Rhizome capabilities of the Serval Project software to share information between devices on the AdHoc mesh network. At the discretion of the user it can also generate JSON files using the GeoJSON format.
These JSON files were used to populate a website to share information with people who were not participating in the exercise. The JSON files were collected using Rhizome by Dr Paul and then uploaded to our server via a satellite connection.
In keeping with open standards and data formats the website used OpenStreetMap in conjunction with OpenLayers to generate the map on the webpage. By the end of the exercise the website looked like this:
Using the data that was collected I undertook development of the Serval Maps Data Manipulator application. The intention of this application is to offer ways of manipulating data that is created by the Serval Maps application.
- Merged all of the individual binary location files into one file for each device. (Serval Maps generates one binary location file per day)
- Merged all of the individual point of interest binary files into one file
- Converted the binary location files into KML files
- Convert the point of interest binary file into a KML file
- Generated an “index” KML file to make it easier to view all of the files in Google Earth
The result of these tasks is a map which looks like the screen captures below:
I’m very pleased with the result and the field testing has been very valuable in providing feedback on the existing application, highlighting areas for improvement and indicating directions for future development.