ETech BlimpBot demo report: Success!

Well, that went better. After last week's disaster, Jordi and I worked hard to make the Etech demo of the Minimum Blimp UAV go smoothly. It was almost a full hour, and the room was a bit smaller, so we were optimistic. But we weren't counting on optimism alone: Jordi had done a lot of work on the blimp since TED to help it handle air currents better, including:

Using the voltage regulators on the Pololu Motor Driver boards rather than a stand-alone voltage regulator.
Adding a separate battery for the motors so high current drains and voltage drops don't risk crashing our Arduino. (This is a short term hack while we get our power management properly fixed and switch to LiPos)
Upgrading the vertical motor to a low voltage N20, which has a better power to thrust ratio than the motor we'd been using
Tweaking the code to give full thrust when needed, which is allowed by the better power management
Other tweaks in the software to adapt to the inertial momentum and laggy physics of a blimp in a moving fluid
The result: success! The range of the IR transceiver is still less than ten feet in a noisy environment like a conference hall, but it nicely follows you around if you hold the beacon and walk around the room and otherwise does what it's supposed to do. Presentation went great, full room, people seemed to love it and video crews interviewed us afterwards. (Links to coverage when they go live)
Next: Minimum Blimp UAV 2.0 with a custom PCB, single LiPo and integrated motor driver chips. And then a kit you can buy and build for less than $100. Stay tuned! [Photo of Jordi credit: Phil Windley]

You need to be a member of diydrones to add comments!

Join diydrones

Comments

Howard Gordon March 6, 2008 at 6:40am

When you design your own beacon, consider including two features:

(1) a set of jumpers by which you can program an ID code to uniquely identify different different beacons (your transmit pattern can then vary from beacon to beacon), and

(2) an external interface such as serial or USB which allows another device to control the transmit pattern of the beacon, effectively providing low bitrate remote control from your host computer to the UAV
3D Robotics

Chris Anderson March 5, 2008 at 10:38pm

One point of clarification: our onboard IR Pololu transceiver is just a passive receiver (actually it occasionally flashes its LEDs because we haven't figured out how to bypass that circuitry entirely when we poll the sensors, but they're off almost all the time). Meanwhile we're just using the ground-based transceivers as transmitters (and because they're on the ground we can run them at any voltage we want). Needless to say, this sort of asymmetrical use is a waste of components, power and money and is not what the Pololu transceivers were meant for. Which is why we'll design custom ground-based beacons and onboard receivers for the next version.
3D Robotics

Chris Anderson March 5, 2008 at 7:22pm

We're still running the Pololu tranceivers off a 9v battery, although they can take up to 16v. First thing is to crank the voltage up to 12v and see what diff that makes. Then, in Version 2 where we switch everything to components on custom PCBs, we'll use high power LEDs and solve the problem once and for all.

The NorthStar system that we use for the Maxium Blimp UAV uses such high-power LEDs and focusing lenses and can do 40+ ft. We don't want to use focusing lens for this one because we need a wide spread, but I'm confident that with the right power and LEDs we can get 20ft for each waypoint beacon. We'll navigate between beacon ranges with the onboard compass in V2.0.
Howard Gordon March 5, 2008 at 7:10pm

I don't think a Vicon system fits Chris' $100 budget.

My sense is that the Pololu IR emitters don't have adequate power, so the receivers are getting washed out by background noise. I think he's okay on the receive side.

Welcome! Future Home of Another Amazing Website Powered By Exabytes
Jack Crossfire March 5, 2008 at 6:32pm

The IR scheme seems to be a mess. The indoor coaxial copters R using ground based vision.

This reply was deleted.