Clustering in image space for place recognition and visual annotations forhuman-robot interaction

The most classical way of attempting to solve the vision-guided navigation problem for autonomous robots corresponds to the use of three-dimensional ( 3-D) geometrical descriptions of the scene; what is known as model-based ap proaches. However, these approaches do not facilitate the user's task becau se they require that geometrically precise models of the 3-D environment be given by the user. In this paper, we propose the use of "annotations" post ed on some type of blackboard or "descriptive" map to facilitate this user- robot interaction. We show that, by using this technique, user commands can be as simple as "go to label 5." To build such a mechanism, new approaches for vision-guided mobile robot na vigation have to be found. We show that this can be achieved by using mixtu re models within an appearance-based paradigm. Mixture models are more usef ul in practice than other pattern recognition methods such as principal com ponent analysis (PCA) or Fisher discriminant analysis (FDA)-also known as l inear discriminant analysis (LDA), because they can represent nonlinear sub spaces. However, given the fact that mixture models are usually learned using the e xpectation-maximization (EM) algorithm which is a gradient ascent technique , the system cannot always converge to a desired final solution, due to the local maxima problem. To resolve this, a genetic version of the EM algorit hm is used. We then show the capabilities of this latest approach on a navi gation task that uses the above described "annotations."