Rapid exploration of unknown areas through dynamic deployment of mobile and stationary sensor nodes
Ferranti, E. and Trigoni, N. and Levene, Mark (2009) Rapid exploration of unknown areas through dynamic deployment of mobile and stationary sensor nodes. Autonomous Agents and Multi-Agent Systems 19 (2), pp. 210-243. ISSN 1387-2532.
When an emergency occurs within a building, it may be initially safer to send autonomous mobile nodes, instead of human responders, to explore the area and identify hazards and victims. Exploring all the area in the minimum amount of time and reporting back interesting findings to the human personnel outside the building is an essential part of rescue operations. Our assumptions are that the area map is unknown, there is no existing network infrastructure, long-range wireless communication is unreliable and nodes are not location-aware. We take into account these limitations, and propose an architecture consisting of both mobile nodes (robots, called agents) and stationary nodes (inexpensive smart devices, called tags). As agents enter the emergency area, they sprinkle tags within the space to label the environment with states. By reading and updating the state of the local tags, agents are able to coordinate indirectly with each other, without relying on direct agent-to-agent communication. In addition, tags wirelessly exchange local information with nearby tags to further assist agents in their exploration task. Our simulation results show that the proposed algorithm, which exploits both tag-to-tag and agent-to-tag communication, outperforms previous algorithms that rely only on agent-to-tag communication.
|Keyword(s) / Subject(s):||Autonomous agents, area exploration, sensor networks, collaboration, tags|
|School:||Birkbeck Schools and Departments > School of Business, Economics & Informatics > Computer Science and Information Systems|
|Research Centre:||Birkbeck Knowledge Lab|
|Date Deposited:||07 Feb 2011 14:07|
|Last Modified:||02 Dec 2016 13:23|
Additional statistics are available via IRStats2.