I can think of a few examples, though they are american projects:
1) MIT CSAIL's M-Blocks
2) Berkeley's Terraswarm
Swarm robotic projects vary a lot depending on their objectives. They can be extremely difficult to implement if you want to start from scratch and implement all electronics, mechanics and hardware yourself. However, you can always start with ready hardware (like the Firebird bot from Nex Robotics (or multiple copies of your own design)). In such a case you can focus more on solving the actual problem rather than debug a lot of hardware.
Generally, swarm problems tend to use differently designed bots together as a team to complete an objective or many copies of the same to do something. In the case of M-Blocks above the objective is motion. Another common problem is to reach a point thru a number of obstacles: walls, movable debris etc using differently capable bots. Or how to map a room/area quickly using N bots/planes equipped with cameras etc.
There have been some previous projects in the institute (as BTP/DDP, not STAB) wherein a set of bots were programmed to follow the leader. You could call this a basic swarm project. People in SysCon do indeed have some swarm projects going on, though I'm not really sure of their exact problem statements.
If you are very adamant about a working prototype, just look at any modern car. Multiple independent controllers and sensors (engines, ECU, safety, electronics, power, windows, proximity sensing etc) working together. Its a non-intuitive example because the communication is wired as opposed to the normal expectation of wireless.