Multi-agent coverage and dispatch

We consider the problem of coverage in Robotic Networks: developing an efficient algorithm which is able to perform a deployment in static obstacle-structured environments is our main idea. The particular aim and tasks we chose are 1. the maximization of local communication: we reformulated the algorithm in order to remove the assumption of a centralized server which collects local information from individual robots and performing the distributed approach as far as possible; 2. the definition of an expansion policy when we have multiple source of robot swarms; 3. the insertion of the possibility to detect small non-convex features in the environment which usually result "blind-spots" for some obstacles, assuming a space structure model for robots (e.g. giving them a round shape) in order to define the limitation of smallest non-convex coverable features; 4. the possibility of event detection and focusing by sending a group of robots.

Autorentext

Marco Fabris studied Automation Engineering and he is now attending a PhD course in Information Engineering. His main research interests are Camera Networks and Non-linear Optimization for Multi-agent Systems.