(joint project Sapienza University of Rome – Sistemi Software Integrati company)
Since April 1st 2012, I am working on a project about “Learning for Multi-Robot Task Allocation” for the Sistemi Software Integrati (SSI) company. The objective is to developed a system in which a team of robots explore an unknown huge environment (outdoor and indoor), that can dynamically change during the exploration mission, to find objects of interest. Robots in order to better explore the environment need to cooperate each other, sharing their acquired knowledge and so on. Due to the big number of parameters, learning is applied both to dynamically assign tasks to the robots and to choose the most adequate exploration strategy. Therefore, this will lead to have an absent or reduced tuning procedure of the coordination algorithm whenever there will be a new unknown environment to explore.
- Multi-Robot Cooperation and Coordination
- Reinforcement Learning
- Self Localization and Mapping (SLAM) Multi-Robot