Human victim detection and stereo-based terrain traversability analysis for behaviour-based robot navigation

This chapter presents three main aspects of the development of a crisis management robot. First, we present an approach for robust victim detection in difficult outdoor conditions. Second, we present an approach where a classification of the terrain in the classes 'traversable' and 'obstacle' is performed using only stereo vision as input data. Lastly, we present behavior-based control architecture, enabling a robot to search for human victims on an incident site, while navigating semi-autonomously, using stereo vision as the main source of sensor information.