Beacon Triangulation     There are many ways to configure a beacon triangulation system that can be used by a robot to calculate it's position, based on beacons placed or positioned at known coordinates.  We have selected a few reference projects (see links below) which most closely approximate the approach we are taking.

This project, the approach to which is outlined below, fulfills our interests in finding an inexpensive way to do the job, while being effective enough to meet navigation demands we see for our projects for the year just ahead.

Once this initial system proves itself, we are very open to more and more sophisticated ways to solve or augment the navigation challenge for mobile robots.  But we will always be challenging ourselves to keep the solutions simple and effective, while keeping the costs down.  Robots that navigate well, but that are not affordable by those who would like to use them, aren't terribly useful at all.

The current mission is to build a beacon triangulation system with the following functionality....

• Three active battery operated ultrasonic beacons and an omni directional (acoustic cone) ultrasonic receiver on the robot

• The ultrasonic beacons will each have an RF receiver, as well

• The robot will have an RF transmitter.

• The ultrasonic beacons and RF receivers and the ultrasonic receiver and RF transmitter on the robot must all have an effective range of up to 30 feet.

• In operation, the robot will ping each beacon individually and sequentially with a coded RF signal.  Each respective beacon will, in turn, return a sonic signal.  The ultrasonic receiver board's microprocessor on board the robot will measure the time of flight of the returning sonic signal.  From the time of flight of the signal from each of the three beacons, and the known coordinates of the beacons, the microprocessor will calculate the robot's position and convey that information to the robot's main processor.  This operation, on all three beacons, must be able to be completed quickly enough so that even if the robot is moving at from 5 to 10 inches/second, there will be no appreciable impact of that movement on the accuracy of position calculation.

• Our desired accuracy for this system is for the position calculation to have an error of no greater than two to three inches within the beacon distance range specified (30 feet).

• In establishing the speed of a sonar signal, as part of determining the robot's distance from a beacon, the reading of an on-board temperature sensor is folded into the computation.  Humidity is also a factor, although the calculation is more sensitive to temperature variations than to humidity variations.

• The robot's receiver will have 12 volts DC available to it.  The beacons will be powered by 9-volt alkaline batteries or any other cost-effective means we find is suitable.

• Every effort is being made to minimize the part costs, while maintaining the effectiveness of the solution developed.

August 28, 2002

Reference Projects:

Boulette (Luxemburg):  http://www.restena.lu/convict/Jeunes/beacon.htm  This is very close to what we are looking for, except it uses IR and not RF as the triggering signal.

Robot localization, Carnegie Mellon University (CMU) (Pittsburgh, PA):  http://www.contrib.andrew.cmu.edu/~rjg/publications/FSR99.pdf  This is an incredible system, overkill for what we need, but has many elements that are applicable.

Robot Localization (Mexico):  http://www.icspat.com/papers/498mfi.pdf  This is a good discussion of the solution we are seeking to implement.