AAC: http://itunes.apple.com/WebObjects/MZStore.woa/wa/viewPodcast?id=330632997
MP3: http://itunes.apple.com/WebObjects/MZStore.woa/wa/viewPodcast?id=330633212
RSS feed
AAC: http://feeds.feedburner.com/diydrones
MP3: http://feeds.feedburner.com/diydronesmp3
Find out what states you track in a 17 state extended kalman filter, what it takes to commercialize a UAV and what hardware object tracking involves.
Reed's profile on DIY Drones: http://diydrones.com/profile/ReedChristiansen
Comments
As long as you don't put me in that group, that also comes across as very preachy and extremely condescending. I've been there much more than you know and I can guarantee much more than you. The last project I lead before I retired was extremely large and complex: delivered on time, under budget and on spec, I have a habit of doing that. It is what I did, took projects from concept to full commercial realisation including all supporting infrastructure, much larger and complex than small UAV businesses.
Might be worth going back and listening to episode 16 of the podcast a lot of what OpenPilot is about is explained there, specifically to run an OpenSource project exactly as a good commercial project would be.
There is a highly dedicated bunch of people working on OpenPilot, the skill set is equal to any I have worked with on any commercial project. The raison d'être for OpenPilot it to produce something of quality that is easy to use for humanitarian reasons, UAVs just for the good they can do without people trying to make their fortune off them.
What's with all the thinking in absolutes also? The people that are working on OpenPilot are all professionals, they are doing for OpenPilot what they do in their day jobs. There is no commercial Vs opensource mentality at all, people are welcome to take OpenPilot as the foundations of a commercial venture if they wish as long as they honour the licenses of the project.
Without getting too off track, my issue was simply the statement that it takes a commercial company to make Open Source projects "reliable". This is an isult to Open Source projects and a gross over generalisation. After all the hard work that has gone in to OpenPilot from some very talented people, I'm not going to let a comment like that slide because it is without basis and is FUD.
For myself, I started out with FlightGear. That is 100% open-source, always has been, always will be. But through the FlightGear project I have made many contacts, and pretty much all my current work has arisen from a FlightGear contact or my FlightGear work in one way or another. In many ways FlightGear is my resume whether I like it now or not or even if I try to tell people it should be. But since FlightGear has had 100's if not 1000's of contributors over the years, that actually works to my advantage and people think more highly of me than they really should.
It's hard to be a purist ... even 100% commercial (open source==bad) type people end up leveraging open source tools all the time ... often without even realizing it. And the 100% open source folks who think any exchange of money is bad (unless it is inbound to themselves) often end up buying more than a few things throughout their lifetimes.
In my view open-source changes the playing field. Commercial people shouldn't fear it. All it means is that you have to pay attention to the world and the market place and you can't sit on your butt with a 10 year old product and expect long term success. It might mean that a commercial company incorporates open-source projects into their larger commercial product because the open-source software/hardware has set a new baseline in the market place. Again, that shouldn't be feared, a business can still be adding value on top of what can be achieved for free or little cost, and people will pay for that.
And Reed makes good points about support and documentation and polish of open-source products. Certainly there are many exceptions that can be pointed out, but often these are lacking or different in the open-source world. A person can get excellent support for an open-source product, but they need to know where and how to ask the questions, and they need to show they are 'for' the community and have tried the obvious things first and put in their time up front. If the open-source community senses a person is lazy, they often ignore them or spit them out the back door and tell them to get lost. An engineer on a deadline might want to be able to pick up the phone and have someone on the other end that will do whatever it takes to work through to a solution.
And then I think Reed tried to express this, but perhaps it's hard to understand until you've actually been there, but developing a working prototype is one thing. Developing a product, making it operational, making it easy, and supporting it is an entirely different beast. A lot of open-source folks never head down that path and don't realize what's involved.
It's clear that whatever anyone says, it is companies like Procerus that set the bar that the open-source community is shooting for.
Curt.
Having a great polished product has nothing to do with:
- being opensource or commercial
- price tag
- sold to military, government or commercial costumers
It has to do with many many factors, but those 3 are certainly not part of the list.
But I see that both of you share a huge passion for UAVs and that certanlly helps.
David, I am looking forward for the first version of Openpilot and i would love to help develop openpilot if i could find a area where my skills could help .I know nothing of electronics, Qt, EKF or DCM. But if you need C#, Java or Perl stuff let me know.
Just makes me more determined to show you exactly the power of Open Source with OpenPilot. The opensource community is more than just a place to "harvest ideas". It might end up giving you a severe ass whopping and then some.
Yes there is an issue with some projects, but don't lump them all together, OpenPilot aims (and is well on track) to beating a lot of the commercial offerings in both quality and final polish.
BTW, running a linear kalman filter at a real fast innerloop speed is a dirty hack for Quad based APs, so the Open Source community that you disparage already beats you there in both design elegance and implementation.
A comment for Chris on simulation, Flightgear or X-Plane is much better than Realflight for simulation, Realflight can not simulate anywhere near what is really required for UAV simulation for example GPS. Flightgear can fly over the actual route you have planned using the real terrain! This is a case where the Open Source community already has the best tools in Flightgear.