 |
|
|
|
|
Giovanni Reina's Homepage Master Student at IDSIA - Istituto Dalle Molle di Studi sull'Intelligenza Artificiale Galleria 2 6928 Manno-Lugano, Switzerland Phone: +41 586 666 715 email: gioreina@idsia.ch _____________________________________________________________________ |
||
MyselfCurrently, I am doing my thesis project at IDSIA - Istituto Dalle Molle di Studi sull'Intelligenza Artificiale on swarm intelligence. I planned to obtain my master degree by Politecnico di Milano on July 2010. Before joining IDSIA I worked on an European project: PHAROS (Platform for searcHing of Audiovisual Resources across Online Spaces) for Microsoft in Oslo (3 months) and for WebModels srl in Como (8 months). I received my Bachelor in Computer Engineering from Politenico di Milano. Como is my hometown, which is situated in the north of Italy. I spent the last year working and increasing my experience on the research field of web technologies. Before terminate the master I am exploring also new branches, working now in Artificial Intelligence field. I have the pleasure to work at IDSIA, a good research institute of Lugano. My project at IDSIARobot Navigation and Path PlanningMy project is part of the the Swarmanoid project (IST-022888), a Future and Emerging Technologies (FET-OPEN) project funded by the European Commission. Swarmanoid project: The main scientific objective of this research project is the design, implementation and control of a novel distributed robotic system. The system will be made up of heterogeneous, dynamically connected, small autonomous robots. Collectively, these robots will form what we call a swarmanoid. The swarmanoid that we intend to build will be comprised of numerous (about 60) autonomous robots of three types: eye-bots, hand-bots, and foot-bots.
The Swarmanoid project is the successor project to the Swarm-bots project, and will build on the results obtained during the Swarm-bots project. My current project at IDSIA concerns robot swarm navigation with the goal of moving a rigid object from an initial location to a final destination through a constrained path made of a sequence of relatively narrow alleys and turns. The project can be divided in two parts:
Distributed path planning In the environment there are some entities (in this case eye-bots, but it is not strictly related to a particular robot architecture) which have a vision of the space from the top, sticked to the ceiling. Partial vision: Each robot as a partial vision of the environment, he can detect the obstacles below him and the position of his neighbors (other robots near him). Noise: The vision and the neighbors positions are affected by error, that is introduced with a zero mean Gaussian distribution. The effect of the noise will be a displacement (roto-translation) in the neighbors maps position. Path planning: In order to calculate the path form the start to the goal position, the robots generate on their map a potential field which attracts the object to the goal and rejects it from the obstacles. The diffusion of the potential starts from the goal and it is spread out through the robots. The path can be outdrawn fairly easily following the decreasing potential. Distributed: The robots have to cooperate and communicate in order to obtain the final path. As first step each one has to detect his neighbors, then the swarm have to spread the potential communicating just some key points values on the frontier with the neighbors. As last step there is the real path calculation, during which, starting from the one near the start, the robots calculate independently their own partial trajectory. When the object exits form the limited view, the robot send the information of the last calculated position to its neighbor, which should continue the process. Fault tolerance: this architecture is robust to failures. It has been developed a wide message schema that allows to the robots to communicate local failures, obstructed narrow passages, loops in the path or unexpected obstacles due to a noising initial map acquisition. Swarm navigation The wheeled robots (in this scenario foot-pack) should coordinate themselves in order to follow the eye-bots path. They have to understand how to move the object and react to any kind of possible error in the path execution. The challenging issues are:
|
|||