Dynamic Waypoint Navigation Using Voronoi Classifier Methods
James L. Overholt, Gregory R. Hudas, Graham Fiorani, Matthew Skalny, Allan Tucker · 2004
This paper details the development of a dynamic waypoint navigation method which introduces and utilizes Voronoi classifiers as the control mechanism for an autonomous mobile robot. A Voronoi diagram may be generated by any finite set of points in a plane. For mobile robot control each point in the plane represents a Voronoi classifier. The classifiers are used to generate Voronoi regions. As a robot comes into a Voronoi region the classifier will act as a control input; providing a new waypoint for the vehicle to follow. The robot moves towards the new waypoint unless interrupted by an obstacle or wall. The robot will get a new waypoint from the classifier in the robot's current Voronoi region. This process continues until the robot has terminated at a desired position (goal) or runs out of power.