Low-Cost Localization for Educational Robotic Platforms via an External Fixed-Position Camera

One of the major limitations with many off-the-shelf, inexpensive robotics platforms is the lack of a simple way to localize the robot. This is an even bigger issue for educational purposes because the cost of a complete solution is usually an important concern. Many great educational projects utilizing a robot become infeasible because of this limitation. This paper attempts to bridge that gap by presenting a method for performing localization in an unknown environment using a single, fixed-position, external camera. The camera’s position does not need to be configured; instead, the localization algorithm will treat image coordinates as features in a topological map. Furthermore, the work is done with an iRobot Roomba vacuum cleaner along with a web-cam to keep overall costs affordable. The low cost of the solution combined with the lack of a complicated configuration requirement helps make the Roomba more useful and accessible to robotics and artificial intelligence education. Results validate the approach and show a significant improvement over relying solely on