Cooperative Perception and World-Model Maintenance in Mobile Navigation Tasks

Cooperative perception and group navigation oriented, world model architecture are discussed. Multiple autonomous agents methodology has been adopted to achieve cooperative behaviour of robots. Special attention has been paid to the distributed perception process. A number of world-model representations has been considered, and the structured semantic net of relationships has been established. Feature extraction and estimation, sensory data-fusion, and world-model maintenance constitute a set of concurrent processes, driven by the team of agents. Task-oriented maps are produced by grouped sensors. Sensed and abstracted features are matched for the consistency check.