Motion optimization scheme for cooperative mobile robots

Abstract

Scene understanding is an essential element involved in the exploration of an unknown environment when using mobile robots. In this regard, scene understanding refers to the identification and localization of elements in the scene. When the scene is dynamic, it has objects that move around and they need to be characterized and discriminated from the static elements. This is essential for the SLAM (Simultaneous Localization And Mapping) process when external global localization information is not available. We present some techniques and algorithms to accomplish these goals with regard to a scenario where we have a co-operative of robots each equipped with a monocular camera. The robots have three degrees of freedom (x,y,θ) and observe a dynamic scene. The objectives are to use the static elements of the scene for self-localization of robots and to estimate the velocity, trajectory and structure of any moving objects with a view to coordinating robot activity. We discuss several metrics for structure estimation that could be used for the co-ordination of robot manoeuvres and develop optimized steering mechanisms, which reduce the errors associated with estimating key parameters. © 2010 Springer-Verlag Berlin Heidelberg.