Our next ArduPilot: the dual-core ArduPilot Pro!

[Chris Anderson]

[NOTE/UPDATE: All the below, from long ago, has been superseded by events. Basic ArduPilot proved capable of every this described here, so we cancelled this version of the product. Thinking about what the next version of ArduPilot could be can be found here. ]

Our main entry-level ArduPilot is designed to be simple, easy to use, and cheap so it just does navigation and leaves stabilization to an stand-alone FMA Co-Pilot with infrared ("thermopile") sensors. But you can tell by our release of our own IR sensor boards the other week, the ultimate aim of this project has been to release a more advanced version that does it all: navigation + stabilization in one. But a standard Arduino isn't powerful enough for all that. So what's the solution?

A dual-core Arduino! So here it is: ArduPilot Pro, with two Arduinos, a MUX/failsafe and built-in GPS onboard. Jordi's design incorporates everything you need for a fully-functioning autopilot, with a target price of under $100, including GPS and thermopile sensors. One ATMega168 processor handles stabilization and the other handles navigation, but because they talk to each other, you get a fully-integrated autopilot, with control over all aircraft channels.

We could have just switched to a much more powerful processor, but that would have cost more, be harder to program and wouldn't benefit from the easy-to-use IDE and the software libraries available for the fantastic open-source Arduino project. The downside of doing it as a dual-core board is that we have to program each Atmega168 processor, as well as the MUX's ATTiny processor, separately, which is why there are three ICSP ports on the board. But in the commercial version, which will come with all processors pre-programmed, there will be little need to fiddle with the stabilization and MUX code, so they can be treated as black-box hardware. (The ICSP ports are just there for anybody who wants to fiddle with them anyway.)

We haven't tested this one yet, so we're not quite ready to provide a link to buy the board. But the Eagle 5 files for the schematic and board are here and here.

[Phil Wilshire]

Hi Chris,
Looks good.

Do you have any code you can share yet ??

Phil

[Phil Wilshire]

Hi again Chris,
What's the plan for an antenna ?
Phil

[Chris Anderson]

Code coming (for the ArduPilot, not the ArduPilot Pro) next week.

The antenna can be any patch antenna with a SMA connector, like this one. I use this one when I'm testing indoors

[Carl Henrikson]

Absolutely fantastic ! when do you expect to be shipping the commercial version ? Will selling of this as a complete product outside of the US , to potentially evil counties (like Sweden or Canada), be a problem? ;)

[Chris Anderson]

Carl,

I think if we market it as an "autopilot development board", like Sparkfun does with theirs (which is very nice but isn't based on open-source hardware and software and costs three times our target price), we should be fine shipping it anywhere. Or at least that strategy seems to work for them...

I don't think we'll be shipping a commercial version of ArduPilot Pro until near the end of the year. But ArduPilot should be fully done, tested and read to sell by October.

Perhaps it goes without saying, but you don't need to wait for us to sell these. As always, we'll provide everything you need, from links to components to code, to make them yourself. This is DIY Drones, after all ;-)

[Spencer Riggs]

When might just the PCB be available? I have no problem with assembling my own boards, and ArduPilot Pro is what I am looking for!

[Chris Anderson]

Octane--

You can take our Eagle files and submit them to any fab you want and get the boards right now. But we want to test them properly before selling them here.

[Richard Embry]

Now if it just had autoland capabilty. possibly integrating the AGL module from micropilot?

[Chris Anderson]

The board is intended to be very expandable. We break out all the unused digital and analog pins and the code is all open, so people can add any sensors and functionality they want.

[hoopty]

Is the idea that they will need IR sensors or can we be free of them