Self-Organized Exploration System: for Unknown, Cluttered and Confined Terrains
Building 22 - Meeting Room
January 26th, 2011
In a disaster scenario the goal of a search and rescue operation is to locate the highest number of victims in the shortest amount of time, while minimizing the risks to rescue workers. Nowadays with advanced technology, Artificial Intelligence are employed as a safer and faster solution to locate the trapped victims inside pancake collapses of a building. This is highly achievable by means of exploration strategies that perform fast but thorough searches of unknown terrains and report back their collected information to the rescue team outside the collapses.
We enhance the performances of such search strategies, applied on mobile agents, by: i) Reducing the positional uncertainty, i.e. the negative effect of transition errors on the information collected on the fly, for the search agents; ii) Reducing search redundancies caused by unnecessary revisiting of explored areas; iii) Preventing congestions inside narrow pathways due to a large number of agents. In this regard we introduce /TeSLiSMA/ a novel self-organized exploration system, in which its autonomous agents are able to speed up the search operation by sequentially partitioning the exploration terrain into subspaces and distribute them among themselves. I will present the superior performance of /TeSLiSMA/ and compare it against other competing search techniques, such as ANT, TREE, Frontier-based and Brick & Mortar.
Advanced software architectures and methodologies