Challenges for autonomous action of mobile robot involve robust control, localisation and mapping. Complex mission can be achieved by a team of robots able to coordinate themselves and reorganize to perceive the environment under various aspects.
Autonomous capabilities of robots in unstructured environments require tools able to manage uncertainties. One of these is the interval analysis already used for state estimation and robust control of underwater robots.
Mobile robots require dedicated sensors, measure understanding and data processing by perception algorithms able to give feedback to action and induce behaviour change if necesary. In real-world, dynamic environment, the information is spatially and temporally distributed and has to be processed and fused to provide a relevant global knowledge representation.