3D Robotics

Sparkfun AVC Registration Now Open


It's that time again! The Sparkfun Autonomous Vehicle Competition registration is now open. Changes for UAVs include separate fixed-wing and rotary-wing categories and extra points for popping bal;oons:

AVC is back and we’ve made some improvements to everyone’s favorite robotic challenge. If you’re not familiar with AVC (or Autonomous Vehicle Competition), it’s SparkFun’s yearly competition where we let you put your autonomous creation through its paces. This year’s competition is going to be June 21st, and will once again be held at the Boulder Reservoir. Check out the recap video from last year.

Registration is now officially open, so head on over to the AVC website and get your robot registered for 2014! Registration has been updated a bit for this year, so make sure you double-check all your information. Also, students teams are encouraged to register! There are special awards just for student teams, so let us know you’re a student team and what school you’re associated with when registering. Although the courses and scoring are remaining largely the same as last year, we’ve made a few minor tweaks to the ground and aerial portions. For a refresher, here are the full rules for 2014. Now let’s talk about the changes we’ve made for this year.


alt text

We’re adding a new feature to the ground course this year that will make the GPS haters (and anyone competing in the Micro/PBR class) happy. We will be adding a line through the course, so you no longer have to rely solely on GPS for navigation. The line will snake through the course, avoiding obstacles, and ultimately end up at the finish line. Coming soon, we will have a course preview video that will show exactly what this line looks like. But for the time being, we can tell you that it will be a white spray chalk line (similar to what’s found on football fields, etc) that will be on the course. It may have some curves, breaks, or other tricks. Generally speaking, it will get harder as the course progresses.


alt text

The first change for the aerial competition will be class designations. This year, there will be two aerial vehicle classes, fixed wing or rotating wing. So quadcopters will be competing only against their own kind; and fixed wings will be competing only against other fixed wing crafts. Not only does this mean you are only competing directly with similar vehicles, but there will be two pools of prizes, so your chance of winning is better!

As we alluded to in our previous AVC post, we needed to make the aerial competition more challenging. When we were planning for 2013, I distinctively remember almost scratching the idea of the wicket or the aerial drop because we thought “no one’s going to try that.” We were dead wrong.

This year we are introducing a new obstacle, the Red Balloons of Death. We will be randomly placing 3 large (roughly 10' in diameter) red balloons somewhere in the aerial course. We’re not telling you where. You are allowed to avoid them, but if you choose to destroy them you will get extra points. You get 25 points for making contact, and an additional 25 points (for each) if you manage to destroy it. Last year was the first year we held the competition over water. Ironically, it was the first year we didn’t give out a water hazard award. This year, I REALLY want to give out this award. If you figure out a way to fly through balloons and still stay up in the air, bring it.

alt text

AVC is always a lot of fun for everyone involved, and we highly encourage you to show up, even if it’s just to watch. The event is free to attend, just make sure you register over at the AVC website as a spectator. At the very least, you get free entrance to the Boulder Reservoir. We once again are not capping the registration for entrants, so if you want to compete, there’s no rush on getting a spot. Just register when you’re ready!

E-mail me when people leave their comments –

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

Join diydrones


  • Here's what we're up against when the balloon pops:

    Check at 0:53:  Latex strips flying like 100 feet.

  • Hi Randy, 

    Odroid and BeagleBone would both be great.

    Clearly the Odroid has sufficient capacity for real image processing stuff.

    Hmm - getting hit in the chest with a ski pole (or the top of the head for that matter) sounds fun.

    Personal reminder - hard hat.

  • Developer

    Yes, not recovering after hitting the balloon is a real danger.  I'm thinking about using an 3dr X8 (which might be better able to survive a hit than a quad because of the extra motors) and then attach some crab like pokers out the front.  Sort of like very short ski poles so the needle bit would be very short and it would have guards so that it can't be buried too deeply into someone's chest in case things go wrong.

  • Randy, yes, I really like the "find the target" aspect.  It's the contact or destroy portion that I think is kinda unreasonable.

  • Developer

    David Anders,

          Yes, it's certainly possible to have a secondary computer direct the copter where to go.  I've started on a wiki page that shows how to connect a RaspberryPi to a Pixhawk.  Once that's working I will probably add a diagram on how to connect the RPi to an APM2.  I may eventually put together a page for other alternative secondary boards like the BeagleBone Black or the Odroid.


          Projectiles aren't allowed it looks like.  I'm actually quite excited about the find-the-balloon component.  Beyond the competition it has some other applications including visual based follow-me and very accurate landings on a target.


  • Developer

    Disregarding all concerns about safety and morality, I would use a camera gimbal with a small camera and the internals of a soft gun aligned so that the pellet would hit at the center of the image at a certain distance. The quad would then just have to do a rough positioning, while the machine visions system would center the balloon in the camera picture and poof!

  • I have really been looking forward to this event for the whole year, but I have to admit to being somewhat underwhelmed.  I was expecting much different challenges and course layout.  The balloon thing really doesn't do it for me.

    I think seeing the balloons is the easy part.  Not easy but... it's been done, so we have some path we can follow.  Physically destroying them with contact from a small flying machine is much harder.  These are not party balloons.  That's the trick that I see, and I don't think I'll even attempt it.  Somebody suggested some kind of spike on a stick, but you'll have a hard time balancing that on a quad, which means it would have to be pretty short, so you're in range of debris when it pops.

    I was thinking of dangling some kinda tether with spikes on it, and just fly over the balloon.  But to have enough mass to be able to actually puncture one of these things, it would need to weigh a bit.  And there's not a lot of capacity for that sort of thing on a 3DR quad or 450 heli.  The way the spectator safety is run (that is to say, absolutely none) basically prohibits the use of machines of any large size.  So constructing some kind of system that can be used on a small copter that could pop one of these balloons, without risk of getting tangled up, or blown away by the air burst...  I dunno.

    I was hoping for challenges more along the lines of more difficult navigation.  Maybe locating a target autonomously.  How cool would it have been if they had an RC boat driving around in the water, and the challenge was to find the boat and drop the tennis ball on it?  How about autonomous pick-up of an object on the island?  So many things could have been done.  But it seems the goal of the event is to try and get somebody to lose their copter in the lake.  I don't get it.

    And then there's airplanes... how on earth is somebody possibly going to be able to pop a balloon with direct physical contact with a foamy?  

    Don't want to be Debbie Downer but, I just don't understand the thinking.  Completely losing your entire aircraft in a lake is not the same thing as busting a wheel off a rover.

  • Developer

    Tracking a roundish object of a known unison color should be relatively easy and fast. No need for fancy approaches like HOG or particle based filters. A quick color space filter should do the trick.

    Here is one such project for the Beagleboard that looks promising.


    Fast Object Tracking - Robot Computer Vision
  • Moderator

    Randy, I've been looking at doing something similar, just not with Pi.  Is it possible to get the Pi to direct the APM where to go?

  • Developer

    Here we go!!!

This reply was deleted.