[Aerotenna] recently announced the first successful flight of an unmanned air vehicle (UAV) powered by a Xilinx Zynq processor running ArduPilot. The Zynq is a dual ARM processor with an onboard FPGA that can offload the processor or provide custom I/O devices. They plan to release their code to their OcPoC (Octagonal Pilot on a Chip) project, an open source initiative that partners with Dronecode, an open source UAV platform.
The [Aerotenna] team used a DJI F550 airframe and plans to test on more hardware, soon. There wasn’t much technical detail in their post, but a post on Xilinx’s Xcell Blog had a bit more detail. In that article, one of the founders of [Aerotenna] says that the FPGA part of the Zynq allows them to do real time processing more efficiently and handle computationally intensive tasks. Flight dynamics and multiple video feeds are mentioned.
The Zynq board runs Linux and the ArduPilot flight control software (which is, of course, open source also). The Zynq is overkill for a lot of projects, but flying a serious UAV is probably a good fit for that much horsepower. We’ve seen a few neat Zynq boards available recently including a hack to turn one into a Parallax Propeller, which–oddly enough–doesn’t have anything to do with flying.
We were surprised that [Aerotenna] didn’t post any video of their flight. However, you might be interested in this year’s keynote for the Embedded Linux Conference where Dronecode’s [Andrew Tridgell] talks about the architecture of the Ardupilot.
This is not the first UAV to fly with a Zync. It might be THEIR first flight, but not THE first flight. You had a story about a UAV ages ago: http://hackaday.com/2014/05/11/phenox-wherein-quadcopters-get-fpgas/
Aw Snap! They’ve been HAD!
More preciously, should be the first flying with Ardupilot based on Zynq, that is the key to make it open source.
As [cheng] points out, I left off one important part: with Ardupilot. Just clarified that. Thanks!
with lots of processing power, it would be interesting to explore using video feeds as inputs to the position sensing system – might be useful for correcting gyro drift etc.
The “mouse sensor” (aka. flow sensor) is mostly what you describe and is for a long time supported by ArduCopter. The Parrot commercial toy quads also use something like that.
Make a 4KM quad with a 1080 or 4k HUD kit and I’ll buy it.. Then probably make it more efficient by using low-pitch big props and switching power reserves.
I don’t really care about FPGAs or firmware it’s a RC flying object you can’t legally fly out of line of sight or a small altitude and even if you could, you couldn’t properly control without autonomous backup navigation firmware..
When we are allowed to make autonomous stuff travel the world then I’ll worry about power-bus, cell tech, autoimmunity algos, and MIPS..
I will say this.. Conservative engineering doesn’t exist even in military tech right now.. Power cells aren’t the problem.. Slapped together library firmware and total absence of power management is..
I’ve always had something like this in mind, but actually my idea (never made it) was to use the current flight controller in my F550 (Naza-M) and place an arduino in between the receiver and the Naza, and fake the receiver signals to move the F-550 around, or upon a controls combination, switch back and bypass from the transmiter, any thougths?
hi im want domething like this(autopilot on zynq) i worked with zynq a lot but i dont know any thing about autopilot how i should start that?
Buy a board, install Vivado, then code a lot. Once you are done, upload it to the zynq device, connect motors and fly.