Hi all Open-Source enthusiast,
Paparazzi UAV, the real Open-Source autopilot driven by a community of users without an economically driven goal. Today we are adding an extra autopilot board to there long list of supported autopilots.
The Pixhawk autopilot board will be integrated in our next software release! The video shows the first flight with Paparazzi UAV on a Pixhawk, more videos will follow.
Why choose Paparazzi UAV?
If you are looking for an advanced, modular open-source autopilot which has features others can only dream of, you choose Paparazzi UAV.
You can make the comparison between Windows and Linux. For example APM is the Windows unfriendly plug and pay software, While Paparazzi UAV is the Linux of autopilot systems, where everything you can imagine is possible.
We make it possible for you to choose the best software for your needs, on the hardware platforms that are right for your project.
Wishing you successful flights,
The Paparazzi dev team
Comments
Dear all,
I hope it became very clear that opinions and statements in this post are not shared by all “paparazzi developers”. I confirm they are not. Moreover I would like to stress that this post (until it is updated) is not even backed by the institutions mentioned in the video either.
I think it’s great that users get many projects to choose from, each with their specific advantages. It would be a very long list to mention all open-source/CC developments everywhere that have inspired other open-source communities lately. So please try to neglect the tone and statements in this post, keep up the positive sharing and based on this: the many great projects and flourishing open-source communities.
Sincerely,
Christophe
PS: Since a few replies seem to ask about the features (*) and difficulties (**) of paparazzi-uav project, I will attempt to make a personal summary about what kind of solution paparazzi is trying to provide:
(*) a project with -- in a single branch -- all the code needed to run on everything from tiny helicopters, over fixed-wings & hybrids, several commercial drones, opti-track or vision-stabilized indoor drones, vision-guided drones to multi-redundant board large UAV, all combined with some recent advances in control, with scripting flight-plans, a growing amount of efficient in-the-loop computer vision concepts, roughly 25 board types not counting clones and versions (incl. naze32, cc3d and in the next release also PX4) with at least over a 100 flight sensors/actuators options + many more as payload.
(**) BUT, as already mentioned by @gautier and @Fnoop (thank you for looking at it): a lot of fast changing options come at a price: it is quite big and hereby has a steep learning curve and is not on Windows (for now). And while there are some videos trying to help you started quickly, with all these options and versions, documentation is still quite "homegrown" indeed.
I hope it nevertheless contributes to all opensource autopilot developments and will be useful for many.
@Fnoop,
Yep, it's a new thing. I have admiration for you for trying to follow the manual. If you
One of the longest going discussions in the RCGroups UAV area is one I did where I did what you did and simply wrote down my experience. Some Paparazzi developers at that time were on RCGroups and jumped in to help me. In the end I got flying and never stopped. The friendships I made are probably made here all the time as people come to start their ArduPilot builds.
Having done Paparazzi so long and helped so many get started I can say some generalities about Paparazzi. In general it is not aimed at a retail consumer/customer. Paparazzi vs retail products will favor the retail products as Paparazzi code, ENAC designed hardware, is created by people who receive no compensation for it. It would be nice but the focus is on features for those using it not features for marketing and sales people to sell it.
Another generality is that the more complex that task is you need done in the air the more Paparazzi will meet your needs. In this way those who have very complex flight needs will eventually find Paparazzi to fit their needs extremely well. I don't mean the sort of thing that wins a race from one or more points back to the starting line but complexity like mimicking the sort of flight found in nature where based on sensors you can alter the flight to meet objectives. The competitions funded by Universities and the Outback Challenge is a good example of complex flight plans. Paparazzi has done well in the OutBack challenge. I helped a grade school class get their hardware and they took first price their initial go at it some years ago.
Reliability measured in actual flight hours flown seems to favor Paparazzi software and hardware. AggieAir has an unprecidented level of detail around their flights from rigidly followed pre-flight, flight and post-flight check lists to the data they keep and log for analysis. Completely objective data that in their experience shows Paparazzi to have a near perfect record in flight. So if you want what you fly to always return this seems to favor Paparazzi and I'm trying to say this in a very objective way as I'm just the messenger of all I've seen in a great many years of reading posts here and other forums. I will say retail customers will probably unfairly skew data of this type as they likely rush through directions and do not keep or follow accurate check lists and logs.
If you get stuck anywhere contact me. Start a build thread here and I'll help if I can. To me it's not us vs everyone else. We are all learning and sharing and any single projects success is a win for everyone if they share how they did it with us all to learn. Be assured my goal is you learn enough to fly both and then be able to simply choose the one that meets your needs best.
It's best to reaize the OP wrote that post likely tired, in another country, with the best of intentions but the translation just fell short of the intended message. Just try and not retranslate back any of it because even more will likely be lost. Dig around the Wiki you'll see a great many ArduPilot and 3DR things there pointing to where to get them. No one would come here to try and say one is better than the other. I see it, just I know that's not what was intended I know those people too well.
@ Tony K We use standard LiPo batteries. They allow 1h flights at the same altitude or 5000m combined altitude in profiles. The most important thing is to not let them cool down. We put them into a 12V cooler box that is set to "heat" and keep them at +40°C. The aircraft battery compartment is air tight. The self heating is enough to keep them at a decent temperature.
@ MArtin M
What type of batteries do you use with temp like that. Do you notice a drop in performance ?
@ Felix Thanks for the points. I actually meant more on the SW side. Why would i fly my pixhawk with paparazzi as suppose to APM ? Thanks again.
Paparazzi sounds interesting and I've spent a day or so reading up and trying to install the software. My first impressions are that it's intended for a completely different audience to APM. After eventually finding most of the information from the (erm.. 'homegrown') website it took me several hours to install the software. It runs on debian/ubuntu (not 'linux' - any rpm or other system is SOTL - I run fedora normally so had to fire up an ubuntu VM) but involves downloading a large toolchain and compiling from source - it doesn't run on windows and I wasn't prepared for the massive download or changes needed for osx. There are no installers or binaries for either the GCS/center or the FC firmware, and that sets the tone for the rest of the project I think. Lots of people who use the likes of APM (and certainly DJI products!) will stop at this point and never give paparazzi a second look back. For those that get past that point it looks pretty cool. I'm looking forward to the pixhawk code so I can actually give it a whirl.
As for this post it's really great that others familiar with the project and one of the developers (Felix) has come out to counter the tone of the OP. Several paparazzi posters have implied that their's is the 'true' opensource project which is just FUD. APM is the same license as Paparazzi (GPL) but px4 on which APM is based/shared is even more 'open-sourcey' with BSD licensing. The hardware licensing is not the same but similar, but the ethos looks to be quite different - Paparazzi hold back open-sourcing the hardware until it's mature unlike apm/pixhawk boards which are usually available even before the hardware is released (eg pixhawk2).
The comparison between windows and linux is one of the oddest (and wrong) analogies I've heard, thanks for the laugh.
What are the 'features others can only dream of'? This is not a leading question, I'm genuinely interested - it's not obvious from the paparazzi website. At first glance the software (GCS) is quite primitive compared to the APM offerings and the APM hardware/firmware seems to be a lot more mature, or at least commercially mature. Paparazzi has been around for longer and has more original hardware iterations but I would hazard a guess and say that APM/Pixhawk has a massively larger user base and thus better testing/bugfixing/release maturity and economies of scale. However if you dig a bit you see paparazzi people doing all sorts of awesome stuff so it's clearly very capable. @Martin Mueller - AWESOME blog!!
Hi Gautier
Thanks for your kind words, may I suggest that a second post is started with the positive spin? I am sure that most of us here are genuinely very interested in what you are doing.
thanks
Phil
As a Paparazzi developer, I'm also very happy that the Pixhawk board is now supported but I don't think it should have been presented like this.
It is actually a good news for both communities. Paparazzi users now have the possibility to use an easily available and widely used board, and Pixhawk users have the possibility to try something different.
And I say different, not better. Paparazzi has been build as an affordable support for scientific research and educational purpose. Since that time, it has been used in many different applications, like professional photogrammetry or just hobby flights. We are trying to build a system as versatile and flexible as possible so that anyone adapt it to its configuration or add it's own piece of code.
Paparazzi is definitely not so easy to use at the beginning but it offers many possibilities and interesting features. I hope some APM or other projects users will give a try, we'll be happy to help them on our mailing list or gitter chat.
@Rob Flying in cold/harsh environments is something that has been done with Paparazzi for a long time. This is a video from 2009, taken over Svalbard at 78°N in (legal) 1500m altitude. The lowest temperature we measured on the ground during that campaign was -35°C.
https://www.youtube.com/watch?v=M1k_TLcQ2ic
And we are still using Paparazzi to do research in Arctic environment. We just came back from one two weeks ago http://blog.pfump.org We also tested quadcopters to allow easier operation by students but we prefer the fixed wing Multiplex Funjet/SUMO for gathering lots of data.
@Rob yeah, I know...
To be be honest, I don't get to fly a lot (or make nice videos)... between my day job and other hobbies this constantly seems to fall down the end of the list somehow ;-)
@Tony: you mean hardware wise (pixhawk compared to other Paparazzi supported boards)?
I don't think it's possible to give a general recommendation here, IMHO it mostly depends on what you want to use it for, how much space you have, if you prefer a casing or not, what hardware you already have, availability, etc...
The pixhawk is a nicely done board (with a two processor architecture), but in the end it provides similar capabilities to many other boards (Lisa/MX, Elle0, Apogee...) in terms of processing power, exposed IO, etc.
It's not a complete coincidence that most projects/boards use similar hardware combinations (e.g. STM32F4, MPUx, MS5611 baro, ...)
For some applications it's nice to have one board with two processors: one for "FlyByWire" (handling actuators and RC) and one for the autopilot ("autonomous") part. A separation like this makes more sense for fixedwings which you can still fully control "manually", less so for multicopters which need active autopilot stabilization...
But if that separation is needed you could also just use two other boards and link them the same way via UART, SPI or CAN.
So I guess this question mostly boils down to simple things like:
Just pick whatever hardware is right for your use-case (or what you already have).
Also I don't have a pixhawk board myself, so no direct experience with it...
-
1
-
2
-
3
-
4
-
5
of 5 Next