Distributed architectures for autonomous underwater vehicles

Focuses on some recent work using techniques from distributed artificial intelligence (DAI) and real time systems to develop a transputer based system which has the potential to implement real time AI decision making with significant computational power in a hardware package small enough to be implemented on a real autonomous underwater vehicles (AUV). The authors are currently using a multi-agent system for work on kinematic guidance and obstacle avoidance for collaborating underwater manipulators, and have the need to take an existing reactive planning system and use it in a real time fashion. The real-time DAI system therefore underpins the eventual implementation of other subsystems for practical AUV application