My latest blog looks at the importance of choosing the right drone software platform to support your business.
Just like Google vs. Apple
When the Apple iPad first appeared on the market in 2010, I didn’t jump in to buy one. I didn’t own an iPhone, I had a company-issued Blackberry, so I wasn’t motivated. Besides, I figured there would be a better model a year or so later. So I waited. By the time Apple released the iPad 2 in 2011 all my friends had one. It looked and felt great in the hand. I thought the user interface (UI) was pretty slick. But I also heard about this thing called Android in development by Google and the Open Handset Alliance (OHA) with a similar and perhaps better UI. I was conflicted about which to buy first. I eventually got an Android tablet on the promise of what could be an open source model. However, after one disappointing experience after another, I got rid of it and switched a year later to an iPad first generation. I stayed on that path and haven’t looked back since.
As Diffen says:
Google’s Android and Apple’s iOS are operating systems that provide a good example of open source vs. proprietary. Both are used primarily in mobile technology, such as smartphones and tablets. Android, which is Linux-based and partly open source, is more PC-like than iOS, in that its interface and basic features are generally more customizable from top to bottom. However, iOS’ uniform design elements are sometimes seen as being more user-friendly.
But wait, I thought we were going to discuss drone software. We are.
For all drones, the interaction between the user and the aircraft, and the aircraft and its hardware is mediated by software. As I have written here, the quality of the pilot experience can be driven by the features and the quality of implementation, but the comparison with tablet and smartphones is a good one. Just as with your smartphone and tablet systems, choosing the wrong software platform for your drone can produce some very high switching costs should you decide later you need to change. In this post, I’m going to look beyond manufacturers’ claims and help you understand the differences with the following explanations of what is it, who makes it, who uses it, and what you need to know.
Read more here: http://droneanalyst.com/2014/05/23/which-is-better-open-source-or-proprietary-drone-software/
Comments
Hi Colin,
I appreciate your analysis but think that a few of the comparisons are misrepresentative.
You attempt to make a general inferred case that currently use of open versus closed software is about equal.
Not even close, there are probably 100 to 1000 open source vehicles / flights for every closed one.
The BIG difference is cost, Boeing, Lockheed, Northrup get paid by the taxpayers to have lots and lots of engineers designing their closed software so they can sell it to governments at inflated costs.
This stuff is completely not affordable at any level by the vast majority of civilian commercial entities who might like to use it.
The companies you mention fall somewhere in between, I am sure many of them doing a great deal more than casual information acquisition from the very open source companies you also mentioned.
I am not accusing them of cheating, I certainly don't have any idea where that boundary is drawn, but a lot of there business comes from what we know.
They actually supply two things of value - direct customer support (good or bad, but at least direct) and the possibility that they can convince certification agencies that their product can be stabilized enough for certification.
The FAA has yet to rule on any of that, but when they do, everybody expects the big bucks government contractors with the big bucks lobbyists to get a leg up (no surprise there).
You don't really have an Apple To Google comparison now either, because in this case Google is a comparative giant and Apple is a couple of tiny little startups with more hype than sales.
The big fish right now in affordable autopilot firmware and mission planner software is us and everyone else is either astronomically overpriced or in a very tiny aquarium (to complete the analogy).
Our open source stuff has been plagued by growing pains, because we have ripped it up and rebuilt it with new stuff before we ever got the old stuff working properly.
But I also believe that we are finally on the cusp of actually getting our act together.
With the upcoming 3.2 copter release I certainly don't expect everything to work perfectly, but for the first time we actually have rebuilt the entire structure to both work in an appropriate fashion and to be readily supportable by numerous developers with a lot less learning pain.
That is huge and it is going to for the first time allow the firmware to be rapidly enhanced and to be more easily debugged.
Basically, except for the unrealistically overpriced government contractors, there just isn't another real game in town other than DIYDrones Open Source.
I know what I am saying is contentious, but there is just too much and too broad a developer base here and now that it is easier to program it is growing.
It is really very hard to see how any closed system can get traction against that.
Best Regards,
Gary
@rob - Thanks for taking the time to read and comment. Q1 Yes. Compatible ESCs for the flight controller and then sometimes the need for recalibration on IRIS. For me the use case that comes to mind is a RedEpic camera being lifted by a large hexacopter (or coaxial) with Pixhawk at the helm. If you know of such an example, I would love to know about. It might be a game changer for those who rely on MikroKopter HL. Q2. For example, swap cameras with different triggers types, power requirements, or even a swap to an analog device.
@Chris –Thanks also for taking the time to read and I appreciate your comments. I have faith that you will get the 3DR and the ecosystem to the level of the current Android tablet experience. I didn’t have room to mention other Open Source providers like Airware.
It’s news to me that PrecisionHawk is using Pixhawk. They still advertise a single CPU as the core offering and are not calling out Pixhawk by name in their Extra Processor ‘options’.
@Seb - Thanks also for reading and your comment. Some sensors and advanced applications do require a whole new embedded processing board – like when you need video processing onboard in real time. From what I read those doing swarm research have used micro-boards that have a full implementation of Linux. I suspect we will see more done in that arena.
Any sensors. If I wanted to strap a geiger counter to my quad to collect data, how would I do that? Maybe that's not what was meant, but it's something I always wondered about. How can I use a sensor to collect data without adding a whole new embedded computer and without messing with any flight code?
Excellent analysis. As with most other open source projects, open source UAV software is almost always the price/performance leader (especially when the price is free). But, as is also the case with other open source projects, that can come at the cost of complexity and a lack of polish if companies and communities don't work hard to avoid that.
The big opportunity in the open source community (and especially the companies built on it, like 3DR) is to add that layer of polish and simplicity on top of the open source core. That's why Android is such a success today (Google caught up to iOS in terms of the user experience).
BTW, some of those commercial projects you mentioned are actually built on open source. For example, PrecisionHawk now uses the Pixhawk platform.
Interesting article and I agree with most of it. But I had a few questions...
What exactly do you mean by this? Are you talking about "which ESC's work with Pixhawk?" Or what exactly? Generally, if you buy an Pixhawk or APM system from 3DR, you get all the components you need that do work together already. No engineering needed.
Which hot-swappable sensors are you referring to? And which components need to be more reliable? For the most part, the Pixhawk uses the same components as the APM with which I assume you are drawing comparison. So how can you insinuate that the Pixhawk is not yet as reliable as the APM?
-
1
-
2
-
3
of 3 Next