Topological Localization

Task Description:

Our approach to the topological localization is as follows: Identify topological and geometric salient features in the environment. Develop control laws such that the robot converges to the features robustly whenever it is in the vicinity of them. Determine a topological representation of the environment in terms of features. Use graph matching techniques to localize the robot. In our work, geometric features are the critical points. We will develop control laws that guide the robot to a critical point whenever it is in the neighborhood of a critical point. We investigate the possible extra features of critical points such as the curvature of the boundary at which the critical point occurs. In this way, we can collect extra information about each critical point so that we can distinguish similar critical points from each other. Even though graph matching may fail under hyper-symmetric environments, we are limiting ourselves to environments which are not hyper-symmetric.

© Copyright 2000 Sensor Based Planning Lab, Carnegie Mellon University. All Rights Reserved.