Way back in 2007, when I was getting started with UAVs, the only autopilot available for less than $15,000 was the UNAV PicoPilot. I got one, stuck it in a plane (see above), and decided I could do better for less. Thus the open source ArduPilot project and the rest is history.
On a whim, I went to see if UNAV was still in business. It isn't. The site is now domain squatted, selling herbal aphrodisiacs (see below). Anybody know what happened to them?
Other companies from the early days that seem to be gone include uThere (Ruby autopilot), AttoPilot, Gluon (Gluonpilot), and perhaps some others I've forgotten. Amazingly, the thermopile-based (IR) FMA stabilization system I first used back then is still for sale (now just $39). One of the reasons we moved quickly into IMU-based autopilots was the FMA had a patent on thermopiles. We told them the thermopile era would end, in part because they were blocking others from using them, but apparently I was wrong ;-)
Comments
It was more interesting when you could match a $15,000 product by doing it yourself, even if the total investment in crash damage made the prices similar. It was a short window when electronics got ahead of industry.
I remember UNAV! We waited a bit longer and got one of the first Atto pilots.
Still got it here too!
Hi All,
When I first got interested in flight controllers, the thermopile was the new big thing and it seemed like magic.
Then we all moved immediately to gyro and accel and thermopile became redundant.
Pretty soon we added GPS and had absolute position - sort of!
GPS has problems, no matter how many channels you put in.
Started to think about thermopile having actual ability to determine the suns position (probably these days bandpass filtered camera better).
How about solar position navigation as backup to GPS - what goes around - comes around.
Just a thought.
The reality is the absolute position flight controller is now pretty much a solved problem, now object relative position is the new high bar, basically discriminating 3D vision and a pile of computing power, Nvidia TX1 for instance.
The future is going to be fun too.
As for patents, I am afraid that the initial concept of protecting the inventors rights has been so eroded and destroyed to defending industrialists rights that it no longer serves a use that is beneficial to the public.
Edison the industrialist with no integrity fought it out with Tesla the scientist with integrity and the scientists lost and the industrialists won.
Edison established the overwhelming precedent that industrialists could take whatever they wanted freely from scientists.
I am sure at least some of you think that is just fine, I don't and never will.
Best,
Gary
Gary yeah my point exactly. Thermopiles have their uses and using them is actually not very hard if your a programmer. I think if I have to I can probably handle doing that myself after all these years hanging out with you guys :)
Neet article anyway Chris good to see you still active and thinking out loud on DIYD. I'm also glad the grumpy interview days are behind us.
Video might well make things seem old fashioned but thermopiles certainly were good with vibrations.
I think video odometry will at some point make the very idea of thermopiles for redundancy seem like a 20th century thing.
I got a big boo from the drones discuss group for mentioning adding the thermopiles back into the Pixhawk code :) There are some good reasons to have it. The one thing I really liked about the thermopiles was you could do a reset in midair without recalibrating. I still think it would be a good back-up system incase a reboot on a complex system as a last ditch effort to save the craft.
Your redundancy would just be the thermopiles and GPS it might not be pretty but you could still RTL or atleast land more safely.
I can't remember if the thermopiles where fast enough for a copter though?
Anyone who has been into sUAS for a while knows of UNAV. They were always the low cost leader - and all that goes along with it. To describe the guy as grumpy seems apt. Maybe curmudgeon would be better. Since I was into rotary-wing I never personally checked out any of their products. I do remember many people complaining of poor performance though. One of the things UNAV blamed poor performance on was weather affecting the GPS. They said clouds could cause positional errors of many dozens of feet or more. This is of course BS, but they wouldn't change their website even after this was pointed out. It was probably too convenient of an excuse for their customers who didn't know better.
As for all autopilots costing more than $15000 in 2007 (except for UNAV) that isn't true. In the interview link in the comments above you can read about the UNAV guy saying he is competing against systems in the $5000-$10000 range. There were even rotary-wing autopilots available then for less than $15000. Off the top of my head I can think of Carvec, Micropilot, and of course DJI who started selling helicopter autopilots in 2007.
Mark, it can certainly go both ways. But I think they they're becoming more of a net negative these days, especially in software. They increase uncertainty and risk, and force reinventing the wheel. Although we have plenty of patents, we have committed to only using them defensively.
It's an interesting thing about patents and innovation.
Sometimes they spur innovation by forcing everybody else to think of something new.
Is this thinking way off base? Can patents actually spur innovation?
-
1
-
2
-
3
of 3 Next