Navio as competition to Pixhawk and knockoffs

I am seeing some knockoffs for PixHawk now for $120.
(http://hobbyking.com/hobbyking/store/__86437__PixFalcon_Micro_PX4_Autopilot_plus_Micro_M8N_GPS_and_Mega_PBD_Power_Module.html).

I am using a Raspberry PI 2 with a C920 to stream video with an APM as the flight controller for <$100 (not counting the C920), compared to your board at $180. Just wondering what’s the advantage of using your board compared to the Pixhawk knockoff + RPi solution. Seems like you are not quite competitive on the price yet. I would buy it if it were around $100 to $120 to compete with the ‘pixfalcon’. Seems like your board is doing exactly the same thing as the pixfalcon, but without the processor, how is it more expensive?

1 Like

NAVIO+ reminds me APM1 (Oilpan) and 2560 CPU board.
Other than the cost competitiveness, I am wondering which is better solution, single high performance CPU board do everything or combination of robust flight controller with companion CPU board for video streaming ?

The main reason I chose a Navio was because the APM trend is “develop till the RAM fills-up then depreciate it”. APM 2 is already obsolete and the Pixhawk is already getting low on memory. Still so many useful features outstanding and the memory is almost depleted.

Looking at 3DR’s direction, following the big DJI, the autopilots are moving to full embedded PCs AND expect another full processor at the ground station/controller.

I saw the PixFalcon and yes for small racing quadcopters it really makes sense as a competitor to Naze and other popular mini boards. It will certainly run for a while, but at which point will the APM developers add so much code that it just doesn’t run on the old (Pixhawk) processor anymore. There are only kilobytes of RAM on that thing.

I know you are talking about driving the PixFalcon with a local connected RasPi for the heavy processing. But that is totally outside the loop and will be limited to MAVLink over a local USB or serial connection.

If people just want that, or something smaller, maybe the Reach and PixFalcon are a better combo (don’t know the price yet though). But I prefer the full capabilities in the smallest package (HAT + RasPi is smaller than PixFalcon + RasPi). Also I’m playing around with getting Windows working with it so I’d prefer to use an autopilot processor board which Microsoft support (really hope they support the Edison).

If Emlid made a Reach+Navio board that would be cool, and VERY small. If Microsoft got Windows running on it I would be even happier. They seem to be pondering it at the moment. Perhaps the next generation Edison would make it.

Anyway, here’s my list of things I prefer about Navio:

  • GPS (main chip) included (and cable cheap).
  • WiFi capability and other USB device possibilities, can’t do that with PixFalcon.
  • Possibility to replace FPV with web cam.
  • Wifi also brings the possibility to eliminate the 3DR radio, within limits (that will suffice for many).
  • Big enough to support whatever 3DR make for their Solo, so when any PixHawk 2 is released (if 3DR still play together with DIY Drones/open source) then hopefully we are future proofed one generation at least.
  • Full speed local access to any of the sensors/ports, for any of your customizations. Possibility to write hooks into HAL source code so back-door to APM also possible.
  • All the other free ports and GPIOs of the RasPi.
4 Likes

Thank you Tony, I couldn’t have said it better myself!

Hi @pduffy – Don’t know if you remember from your threads over on DIYDrones…But you were helping me work on my gstreamer setup similar to yours. I recently acquired a Navio+ board to see if I could scale down the power/networking requirements on my copter. It just seems more efficient and space saving to have all the functions through one unit. I’m still only at the point of doing hover tests but so far I am pleased with Navio+ I’ll begin the gstreamer/picamera testing later this week but I already know I am going to shield it as it is noisy. Also, the Emlid community here is great and very helpful. Your point about the price is something I agree with, but my hope is that as this platform gains popularity the prices will come down.

APM2 = discontinued (so we are wasting time in this discussion)

Pixhawk = very good board but with hw limitations for the future. (So we are wasting time in this discussion too).

Navio+ and Raspberry Pi2 = very good solution for now and the future, very high processor capacity, memory as well, and with Linux S.O. you can do everything you want.

I started my second project with Navio+ last weekend, a very high end project; as soon as possible I will share.

By the way, I flew a lot yesterday, amazing board.

2 Likes

Just a bit off topic but I’m too curious, Why do you want Microsoft Windows on an autopilot? What is wrong with Linux?

That could run into a long discussion :sleeping: The main reason is I see a demand for plug-and-play, simple solutions. Not everyone is a developer or wants the hassle to discover and work with cryptic commands and text files.

Even though Microsoft did a pretty good job of annoying all their loyal fans with Windows 8.x, as a developer I’m still with them, especially for hardware, because I think they still (for now :watch:) have something special:

Look how Microsoft have achieved BOTH hardware and application level binary interfaces, which work right back with code over 10 years old. Apparently the guy who wrote Linux explicitly doesn’t want a binary interface for drivers, seems paranoid about open source. That decision alone kills any chance of true plug-and-play with Linux.

I think Linux is great when your hardware platform is fixed or has a limited set of devices. It used to be the only feasible mainstream system which could run on small devices. But that old argument is rapidly fading, with the RasPi2 and the like on the market, full operating systems can also run on tiny cheap PCs now. So I think Windows IoT has a very good chance.

The other reason I like it is, well I like good architecture, and Windows and all related tools are just cleaner. Linux feels “quick and dirty” to me, but not actually “quick” in the end because everything is so overcomplicated. Ubuntu is better, but still far off.

Of course right now, any hardware company would be mad not to support Linux on an embedded solution. I think it will be a long time and a wish before people will serious consider Windows IoT primarily. It could all just be another flop like the Windows Tablet PC (prompted the iPad) and CE (prompted the iPhone).

Let’s see if the thing flies, that may show some opportunity :smile: If Microsoft ditch IoT after one or two versions then I’ll be programming Linux or whatever Apple’s “new invention” (copy of Windows IoT) is :green_apple: :helicopter:

4 Likes

Old thread but thought I would add to it. with RPI’s you can run cross communication over ethernet with each other and pass huge amounts of data in a timely fashion with extremely low latency. This allows you to have multiple pi’s runing different aspects of your project and feed information into the main FC in extremely low latency situations. I choose to go with a Navio setup for what I am doing as I will required the ability to do real time image recognition and feed data to the fc based on what is found. Also the ability to have real time image recognition with geotagged information from more then one gps source. Also being that it is linux and python has limitless abilities plus your data storage is essentially unlimited it was a clear cut solution for what I am trying to achieve.

There are many makes, breeds of drone/UAV platforms out there. The only issue I have found is none of them really do everything you need it to do. The reason why is lack of processing power. I am eagerly awaiting my hardware delivery to start really diving into what I am trying to accomplish. So far I have invested under $500cad and have all the electronics required less esc/servos/motors/ receivers & Transmitters.

You also have to realise with a rpi you can run any sensors on the market for pis, arduino and more then feed the data from them into the fc for whatever you need it to do. It the open possibilities that made my choice easy.

2 Likes