libprop

libprop uses high-resolution USGS terrain data to accurately predict radio propagation. It's a set of C++ classes that you can easily incorporate into an existing project.

All libprop materials are released under GPLv3 or GFDL.

Online Mashup

We wrote a quick mash-up that uses libprop to show tower propagation on Google Maps. Drag and move the tower to see an updated propagation pattern.

Download

You can download a recent tarball or connect directly to the development subversion repository.

libprop-0.12.tar.bz2 on February 23, 2008

Subversion

Most development work happens in subversion. If you'd like write access, send us an E-mail at .

svn co http://libprop.jsharkey.org/libprop/ libprop

Documentation

Doxygen documentation is provided online and along with the source code.

Mailing List

If you have development questions or are interested in project updates, join our mailing list.

Background

Radio propagation analysis describes how radio signals move through a given area. Given a set of known radio equipment and two specific locations, we can model the environmental conditions between those points to determine if successful radio communication is possible.

libprop was written by Jeffrey Sharkey while a Graduate Research Assistant at the Western Transportation Institute at Montana State University.

Data Sources

The USGS is an excellent source of digitized terrain data, and they continue to improve data quality and resolution as technology allows. Currently they are offering 10-meter resolution data for the entire United States, which is what libprop usually expects.

libprop can accept data in either the GridFloat or BIL (integer) formats from USGS, and it understands native lat/long mapping and Albers projection. For terrain data we recommend using "National Elevation Dataset (NED) 1/9 Arc Second" data source, which provides 10-meter resolution. We also recommend the "Landfire Existing Vegetation Cover" and "Landfire Existing Vegetation Height" data sources which provides vegetation type and height information. Optionally, we can accept the "National Land Cover Dataset 2001" data source which provides land-use classifications. Our algorithm doesn't use this data by default.

Examples

You should consider including a 10dBm link margin to account for other interference.

Path loss between two points

Here is a quick example of how to find the path loss between two points: 

// open elevation data source
SourceGroup* sg = new SourceGroup();
sg->add(new SourceGridFloat(new Convert(), TYPE_ELEV, "data/80214271.elev/80214271.hdr", false));

// define points to test, both with 5-meter towers
Point* start = new Point(45.52391667, -111.2476944, 5);
Point* end = new Point(45.516652, -111.254980, 5);

// assume 10-meter detail, 4-watt transmitter,
// 0dB antennas, and a 900MHz system
// returned value is signal strength in dBm
double loss = pathLoss(start, end, sg, 0.010, 4000, 0, 900);

// link is possible if loss is above the receiver sensitivity
bool possible = (loss > -85);

Testing a roadway

Here is another example of testing all points along an entire roadway loaded from a file. The input file is a series of lat/long tab-separated pairs which could be derived from GIS files provided by a state DOT. We correctly space the discrete points, regardless of spacing used in input file. 

// open elevation data source
SourceGroup* sg = new SourceGroup();
sg->add(new SourceGridFloat(new Convert(), TYPE_ELEV, "data/80214271.elev/80214271.hdr", true));

// define tower point we are interested in with 10 meter height
Point* tower = new Point(45.52391667, -111.2476944, 10);

// load roadway from input file
// split into discrete points every 50 meters
RegionLine* road = new RegionLine("data/mt199-highway.txt");
vector<Point*> list = road->discrete(0.050);

// run a path loss from tower to each road point
// assume road point antennas are all 3 meters high
vector<Point*>::iterator it;
for(it = list.begin(); it != list.end(); it++) {
    Point* r = *it;
    r->towerHeight = 3;

    // assume 25-meter detail, 1-watt transmitter,
    // 10dB antennas on both ends, and a 2.4GHz system
    // returned value is signal strength in dBm
    double loss = pathLoss(tower, r, sg, 0.025, 1000, 10+10, 2400);

    // output any signal strength if not completely blocked
    if(loss == DENIED) continue;
    cout << r->lat << "\t" << r->lon << "\t" << loss << endl;
}

Testing a grid region with vegetation

Here is another example of testing all points in a grid around a given tower, while also adding vegetation information. 

// open elevation data source, along with vegetation information
SourceGroup* sg = new SourceGroup();
sg->add(new SourceGridFloat(new Convert(), TYPE_ELEV, "data/80214271.elev/80214271.hdr", true));
sg->add(new SourceGridFloat(new ConvertAlbers(), TYPE_VEGHEIGHT, "data/22276103.height/22276103.hdr", true));
sg->add(new SourceGridFloat(new ConvertAlbers(), TYPE_VEGTYPE, "data/22273282.type/22273282.hdr", true));

// define tower point we are interested in with 10 meter height
Point* tower = new Point(45.52391667, -111.2476944, 10);

// define area with 5-km radius from tower
// split into discrete points every 100 meters
RegionArea* area = new RegionArea(tower, 5);
vector<Point*> list = area->discrete(0.100);

// test loss between points as shown above
// ...

Future work