Exploring an unknown environment comes with risks and complications, and in some cases an environment may be too dangerous for humans to explore, but immediate exploration is critical, as in the aftermath of an earthquake. Robots, however, can be deployed to seek out points of interest and report back to the waiting human operators. One aspect of a disaster scenario is that communication is often more limited than we are accustomed to in everyday life, so these robots cannot rely on having constant contact with the outside world, or even with all other robots in the environment. In this paper, we present two algorithms for a small team of robots to explore an unknown environment, and use both simulation and experiments with physical robots to demonstrate the algorithms' performance. We provide proofs of correctness and guarantee full coverage of the environment, even with attrition.
展开▼
