December 13, 2016

Don't Buy Another Drone—Build One Yourself

MIT Custom Built Drone
MIT's new interactive program will let you design and build your own custom drone.

Do you love the idea of aerial cinematography, but haven't been able to find an UAV that meets your specific needs? Students at MIT's CSAIL (Computer Science & Artificial Intelligence Laboratory) program might have come to the rescue, developing a first-of-its-kind interactive system that enables you to design, test, and build your own drones—even without a technical background.

For users with very specific and unique needs (for payload, application, form factor, etc.) and no training in computer science, the CSAIL system will make the impossible possible by simplifying the design process and guaranteeing a successful build. According to PhD student Tau Du, “Developing multicopters like these that are actually flyable involves a lot of trial-and-error, tweaking the balance between all the propellers and rotors...It would be more or less impossible for an amateur user, especially one without any computer-science background.”

So how, then, will the program allow the average Joe or Jane to take their idea for a drone from concept to reality? According to MIT News, Joe or Jane begins by choosing from a database of specific parts. Then, the program collects information about their needs in areas like payload and battery usage, and compares their design concept with data about the specified parts and the various physics required to keep the drone aloft. Finally, they report their findings back to the Joe or Jane in the form of a simulated model, allowing them to continue altering their design until they've got an airworthy drone. If you want to get extra geeky, you can read the students' full report on their methodology here.

The system is currently only in the prototype stage, but the intention is for it to become available as open-source. It seems like a revolutionary way to create, and will certainly prove valuable for many drone applications—one of which might be filmmakers designing drones around certain camera kits or shot applications rather than waiting for drone manufacturers to create aerial platforms that fit our specific parameters.     

Featured image: Jason Dorfman/MIT CSAIL

Your Comment

7 Comments

Love the camera on that..........................with that amazing gimbal..................... and the, atleast, 6 propellars that gives it redundancy so I can fly over people...........................

December 13, 2016 at 4:41PM

5
Reply
avatar
Torben Greve
Cinematographer
1002

Coming from building multiple UAV's hundreds of times I feel so much safer going with DJI or FreeFly based on the RD those companies can do. Getting a drone to fly is the easy, keeping it in the air is the hard part. I can understand special need projects but do not involve any people in these special projects. Once you've had a scare with one of these drones you will understand the potential danger. All though disappointed with the number of competition DJI created for the Aerial Cinematographer, I was happy with the reliability they put in their product (mostly just recently, 2014 with the s900 and on). And if it took DJI that long to get a reliable system think about how much RD these one off's will need. Just be safe and don't think that these one off projects will be safe to use on set around people.

December 13, 2016 at 5:04PM

8
Reply
asa martinez
Camera Movement Tech, Camera Operator
101

Don't believe their lies, tried building a drone with a buddy and it just would never fly. Spent more money on it than just buying a properly made one straight up.

December 13, 2016 at 6:38PM

0
Reply
avatar
Tim Brennan
Big Boss
157

they told me I could build a helicopter. the fools

December 14, 2016 at 1:55AM

0
Reply

And then were supposed to code assisting software on our own?

December 14, 2016 at 7:38AM, Edited December 14, 7:38AM

0
Reply
avatar
Adrian Maurud
Cinematographer
158

This is a BAD idea.

Before the multi-rotor copter burst onto the scene - via an open source project that ultimately evolved into Mikrokopter about six years ago - there were only traditional single-rotor helicopters to carry cameras. These required some pretty specific mechanical knowledge and flying skills. Electronics were largely limited to the radio control components, with perhaps speed governors and tail gyros to ease the pilot workload.

In other words, you had to know what you were doing.

Then along comes the multi-rotor, or drone. Full of gyros, accelerometers and other miniaturised electronics. It was still home-build but there were a few kits around to at least help with the individual components selection until DJI, who had previously made only add-on auto-pilot electronic systems for single rotor helicopters, released their S800 hexacopter.

With a very modular design that included clip-on/off booms, plus their own multi-copter flight control that drew on their previous helicopter auto-pilot experience that came in nice neat little plug-together boxes, instead of circuit boards that had to be completed by soldering on a few microscopic little bits, the S800 was quite a step forward.

Unfortunately, DJI had not really ironed out all the bugs before starting to pocket everyone's cash and probably 50% of all those S800's crashed due to unforgivable faults in various components. In the end pretty much every single important component of the S800 was upgraded at least once, sometimes twice - at the customer's expense. This included the booms, the GPS puck, the IMU, the undercarriage and an anti-vibration transmission system that finally isolated the camera from airframe vibrations.

It was an expensive business getting an S800 up to scratch. At the same time as the S800 was released, DJI announced the Zenmuse brushless gimbals. These were ground breaking camera mounts at a time when everyone was struggling with servos and Piclocs in valiant but ultimately vain attempts to get steady in-camera footage.

The BIG move came with the Phantom quad-copter. I can barely bring myself to write the word in this post, such is my disgust for that very capable bloody insect.

Every wanabee twit who knew nothing about everything bought a Phantom and started being a tit in the air. Crappy GoPro videos started flooding YouTube and governments around the world started freaking out about bloody quad-copters (read DJI Phantom) being flown in every stupid inappropriate place by ignorant, irresponsible A-Holes. A-Holes who knew nothing more technical than how to hold down the power button to turn it on.

Fortunately, DJI helped out quite a bit here. Their crappy Flight Control made sure that a large percentage of Phantoms flew off into oblivion - although this did also have the negative effect of freaking out the regulators even more.

Nothing much has changed in the meantime. DJI is pumping out ever more sophisticated and smaller products and wanabees are still dreaming about being 'Aerial Cinematographers', but the general level of knowledge of how these things work, how to analyse and fix faults, how to even fly them without GPS assistance, all this is still dark, foggy science to the vast, vast majority of 'drone operators'.

So for MIT to jump back in time to produce a home-build kit is to invite swathes more ignoramii into the game.

Just forget the idea entirely and let the likes of DJI continue to rule the roost, because their products do actually work now.

If someone wants to build their own drone, let them get practical experience first via the long way round. Because then, and ONLY THEN, will they have any appreciation whatsoever of how these things work and maybe they fill be safer in the air. It is very important that a pilot knows how his aircraft works.

All these cheap, shortcut, easy solutions for the stupid, the ignorant and the lazy are just dangerous.

December 17, 2016 at 3:27AM, Edited December 17, 3:35AM

1
Reply
avatar
Graham HAY
Managing Director, Helicam International Ltd.
177

Building the drone is easy. Trusting the drone to stay in the sky is impossible. Especially when it doesn't half the time for random reasons. Then don't even think about trying to live stream video back. You'll spend hundreds of dollars and hours splicing connectors for different image transmission systems and connectors. Not worth it. Built my own 6 prop and flew a DSLR and RED on it. Not worth it...

December 21, 2016 at 9:41PM, Edited December 21, 9:41PM

9
Reply
avatar
Clark McCauley
Spaceman
2010