Research Paper #552
|Using Motor Actions for Location Recognition
|In Proceedings of 1st European Conference on Artificial Life (ECAL 91), MIT Press, Cambridge, Mass. and London, England.
|We present a Behaviour-based mobile robot that is able to autonomously build internal representations of its environment and use these for location recognition. This is done by a process of self-organisation, no explicit world-model is given. The robot uses an artificial neural network for mapbuilding and additional adaptive processes to achieve successful location recognition. The robot reliably recognises locations in the world and can cope with the noise inherent to the real world.
|NO ONLINE COPY