Previously, I discussed my school project to create swarming robots, and potential tasks for them. This time I’ll talk about some of the hardware, beginning with the control system.

When first concieving of this project, we figured it would be easy to build a robot that could do what we needed it to do. We though a simple microcontroler would do the trick. Sure, that could make a robot, but the trick here is communication. Each robot needs to communicate with every other robot simultaniously. Anything from their current status to actually attempting to querey for data or providing instructions. The details of what would be sent where not important.

We did a bunch of research on what we could use. We looked at GumStix, which are really small embedded linux machines, but they where about $100 a piece, which would go over our budget quickly. We looked at external devices, but nothing seemed well suited for wireless communication like this, atleast nothing that was cheap enough. Wifi adapters all used USB and where expensive, bluetooth was the same deal.

But then I discovered the Wifi car. The Wifi car is a modded RC truck that someone had hacked a Linksys WRT54G into to control via a computer that had ssh’d to the router to run his custom software in the firmware. The router controled the moters and the software had all the controls that where needed to drive the car.

If you sit and think, you will realize the implications of this. For $30 or so, you can find an old WRT54G router, change up the firmware to run linux and do whatever you want, and have Wifi communication, which allows greater distance and greater flexibility. We can make use of pre-existing network stacks which prevent data collisions, and still have an inexpensive vehicle.

Currently, we have 2 routers to hack, one more on the way, and possibly 2 more available to us. We need to play with battery power, and how to interface with a car at this point.

Sorry it’s been a while. Busy busy!

As previously discussed, I am working on a school project to create some swarming robots. In this post, I’ll discuss what those robot might do.

When you think about swarming robots, you might think something like the nanobots from Michael Crichton’s Prey, where thousands of self replicating robots could form up into larger shapes to accomplish tasks. You might think along the lines of the Replicators from Stargate SG-1, where small building blocks could form up and self-replicate and form mechanical bugs. I’d like to say here and now, that that is not what this project is, although some of the same techniques might be used in communication.

The idea here is to have a handful of independent robots that can communicate with each other in order to accomplish some task, or perform some action. This task or action doesn’t have to be complicated, but it has to be coordinated to some degree.

We have discussed several possibilities of what we could have the robots do. First, we though about just having them line up into a formation. But with 5 robots, you’d have a limited number of possibilities, and it wouldn’t be all that interesting, although such an idea is a fundamental part of larger swarms such as in Sci-fi. Another idea was to have them play follow the leader. We could have them find each other and form a line then have the one that was in the front drive some pre-programmed course. That is much more interesting, but easily faked.

We went and looked up what exiting research project where doing with their robots swarms. Most of them used very expensive, but much smaller and more agail, robots than what we have in mind. They would have about 20 of them in a field, and designate some red, some blue, and some black. The red ones would drive around randomly, but turn any others that they ran into to be red. The blue ones would drive randomly, but cause ones that they ran into to freeze. The black ones would just drive aimlessly, I think unfreezing ones they hit. Basically and over complicated game of tag.

We liked this idea, and think that we will have our robots play normal tag, with one being “It” and have it chase other robots, which would try and stay away from It. The It status would pass off when the current It has tagged another robot.

We like this idea as it simulates many different kinds of behavior. There are places where the code has to find other robots like they would to join up (the It robot seeking out other robots), places where they need to stay away, pieces for finding location of yourself and others, and opportunity for the robots to work together sharing their information on the location of the It robot.

Let me know what you think and if you have any other ideas of what they could do.