Working with Mini-ITX-related products throughout the day can be, at times, a static endeavor for me. I take pictures of the still, quiet forms; edit and size them without any of the objects voicing complaints; write little blurbs about them without asking for their permissions; and finally, post the products online where they become virtual representations of inanimate, cold pieces of metal and plastic. It all sounds very sad, but keep in mind here, I primarily photograph people.

However, what I find to be very interesting is when bright, young (or old) people out there take these still, quiet objects and apply them to a task (and almost all our customers do). Then, what really boosts my day is when I get those few chances to administer futuristic terms like autonomous flying robots as I describe one of our customer’s Mini-ITX project. It adds a neat sci-fi edge to my career. And, the application I am about to discuss is certainly teetering on the brink of a sci-fi-themed movie.

Robot Helicopter in FlightLogic Supply donated an MSI 945GM1 Mini-ITX mainboard, 2GB IDE flash module, and an M2-ATX automotive power supply to the IEEE (Institute of Electrical and Electronics Engineers, Inc.) student branch at Purdue University in Indiana. This team of college students (part of the Aerial Robotics Committee) has been working on developing an unmanned robot helicopter that can accomplish a set number of tasks in a designated time frame to compete in the Association for Unmanned Vehicle Systems International’s (AUVSI) International Aerial Robotics Competition (IARC). I have been in contact with one of the brains behind the operation to gather information on the project and learn a little more about how Mini-ITX has assisted the group in achieving its goals.

First things first: the platform

Innards of the Robot HelicopterAs I briefly touched on earlier, their platform consists of the following: an MSI 945GM1 (MS-7265) Core 2 Duo Mobile (Merom) Mini-ITX mainboard, a Core 2 Duo Mobile 1.8GHz processor, 2GB IDE direct-plug flash module, 1GB DDR2 667 RAM, and an M2-ATX automotive intelligent power supply. The Indiana boys found their way to the Logic Supply Web site after revisiting some of their earlier helicopter prototypes. They decided they might need to harness more power and joked about the possibility of having a Core 2 Duo processor onboard the helicopter. After looking around, they found our Web site and made some phone calls.

Falkor in FlightThe helicopter isn’t very big. I estimate it to be roughly 2 ft tall and about 5 ft long. It is commanded by an autopilot, which the team developed from scratch (other teams in the IARC usually buy their autopilots, costing them thousands of dollars even though they can be somewhat unreliable). The computer’s OS is a stripped down version of the Linux slackware distribution. They plan on transitioning to a custom x64-based installation that will alleviate stress placed on the solid state hard drive, and, at the same time, speed up access to critical hard disk memory. The students write all their own software and flight programs for the robot, which seems pretty complicated if you think about it. Flying a helicopter appears fairly difficult. There is much to take into consideration—one being wind. So, if you are not a helicopter pilot, imagine trying to tell a computer how to be one.

Next: the project

The IEEE student branch at Purdue University had a division that was interested in aerial robotics. After some organizing, the newly formed team spent a lot of time researching and experimenting with different platforms so they could build a system that would meet the requirements of the IARC. July 2007 was the first competition in which the team participated. They met the Joint Architecture for Unmanned Systems (JAUS) challenge—a set of standardized messages for controlling unmanned vehicles operating under a single set of instructions. Being the team’s first year, it did not continue onto the next phase of the competition (I’ll get into the competition a little later). Apparently, though, just meeting the JAUS challenge is an accomplishment in and of itself. Many of the competing teams have been going to IARC for years, but no one as of yet has actually won. Pretty tough challenge if you ask me. So that brings me to the next step…

Finally: the competition

The AUVSI’s IARC takes place in Fort Benning, Georgia. This past year’s event was the 17th annual competition for the IARC. Each year, students from colleges and universities around the globe gather in Georgia and attempt to meet all 4 levels of the aerial robotics challenge. Level 1: students must construct completely autonomous flying robots that go through a 3 km entryway to a designated area. Level 2: the helicopter must then identify a particular building and the possible openings to the building. Level 3: the helicopter must select an opening in the facility and enter it, or, instead, it can send an autonomous probe into the facility. Either one of these objects has to search for a particular target, find it, then transmit pictures 3 km back to the starting point. Level 4: all these previous tasks are combined into one grand mission and the pressure is on.  The automatons are now in a timed race to see which one can seamlessly complete the entire task in less than 15 minutes.

Over the years, the IARC has added different mission levels to its program.  When it first started, there were fewer levels, and their difficulty wasn’t as great due to longer time limits. So, there were some contest winners. But now, no one has completed all 4 levels for about 8 years. The money pot continues to grow (it is about an $80,000 prize), and more student groups are accepting the challenge.

The role of Mini-ITX

There is obviously limited space onboard this helicopter. So, opting for a smaller form factor was key in the overall design. There also is a lot of vibration, noise, and electromagnetic interference onboard this helicopters, which is why the group chose a solid state platform. The platform also needed to withstand a wide temperature range and be robust—the experimental process is somewhat timely and there is a lot of trial and error, so the board needs to be able to offer some options.

Overall, I am pretty impressed with the whole project, but I am surprised that the military doesn’t already have legions of robot helicopters. This puts me somehat at ease; at least I know that we have more time before the Terminator movies become a reality.

If anyone is looking for more information, please contact me direct, and I will put you in touch with the group.