Big news from Dronecode, where Qualcomm, whose Snapdragon Flight board was the hit of CES, has moved up to become a Platinum member. Also another 27 companies have joined.
SAN FRANCISCO – January 12, 2016 – Dronecode, the nonprofit organization developing a common, shared open source platform for Unmanned Aerial Vehicles (UAVs), today announced major milestones that include investments from 27 new member organizations and the formation of technical working groups to advance the Dronecode open source platform for UAVs.
With investments from more than 50 member organizations, Dronecode has nearly tripled membership since its formation. One of the founding members, Qualcomm Technologies, Inc., a subsidiary of Qualcomm Incorporated, has increased its investment to the platinum level as it introduces new products for the UAV ecosystem. Qualcomm Technologies, Inc. recently launched Qualcomm® Snapdragon Flight™, a highly integrated 58x40mm board targeted specifically for consumer drones and robotics applications. Snapdragon Flight is based on a Qualcomm® Snapdragon™ 801 processor, along with Dronecode PX4 open source UAV software and robust connectivity, bringing cutting-edge mobile technologies to create a new class of consumer drones.
The highest growth rate for drones or UAVs is commercial opportunities across applications in areas such as agriculture, energy, utilities, mining, construction and more (BI Intelligence’s “Drone Market Forecast 2015”). By uniting the industry around a neutral, open source platform, the Dronecode Project aims to help the drone industry overcome software and hardware barriers to accelerate drone adoption for these critical applications.
The newest Dronecode members include silver members Aerotenna LLC, AirMap, Airphrame, Altitude Angel, AutoModality, BirdsEyeView Aerobotics, CUAV, Droidika, Dig.y.Sol, DroneDeploy, DroneWorks Inc., Emlid, EnRoute, Falcon Unmanned, Hex Technologies Limited, Incite Focus, Inspectools, Matternet, ProfiCNC, Sentera, Skedans, Yin Yan Tech US Inc. and Zubax. New sponsored members include Humanitarian UAV network, OpenTX Project, Stanford University Aerospace Design Lab and UAVCAN.
The Dronecode Project is also forming three technical working groups to ensure better standardization and interoperability in the areas of camera and gimbal controls; airspace management; and hardware/software interfaces. Details include:
- MAVlink Camera Working Group: aims to assist camera manufacturers implementing the MAVlink protocol in cameras. The group will also work with developers and manufacturers to expand the Dronecode platform so that it can support additional cameras and functions.
- Airspace Working Group: formed to establish common data types, units and formats that all airspace providers can transmit and receive. The working group will also lead the discussion on best practices for how to ensure separation between aircraft(s) and establish agreement on common response behavior.
- Hardware Working Group: the mandate of the hardware working group is to establish mechanical and electrical standards for interfaces to the autopilot and the peripherals. This will create a more formal interface between hardware and software development and unite efforts between Dronecode members and the open source developer community working to advance UAVs.
“From increasing member investments to a growing, vibrant developer community, the Dronecode Project’s first year has been extremely exciting,” said Chris Anderson, Dronecode Board of Directors chairman. “By bringing efforts together to establish a common platform and utilizing open source best practices, we're able to build the foundation for a new era of drone applications that extend from the camera to the cloud. The Dronecode "full-stack" platform approach, combined with the hardware and software innovations of its members, will bring about a new generation of drones that are autonomous, aware of their environments, and continuously connected -- an airborne Internet of Things."
The Dronecode Project is a Linux Foundation Collaborative Project. Founded in 2000, The Linux Foundation today provides tools, training and events to scale any open source project, which together deliver an economic impact not achievable by any one company. More information can be found atwww.linuxfoundation.org.
About Dronecode Project
Dronecode Foundation is a nonprofit organization working on a common, shared open source platform for Unmanned Aerial Vehicles (UAVs). Dronecode brings together existing and future open source UAV software projects, including the APM UAV software platform originally developed by 3DRobotics. More than 1,200 developers are working on Dronecode’s six projects focused on maximizing adoption of the project's cost-effective, reliable and technologically advanced UAV software. A Linux Foundation Collaborative Project, Dronecode is an independently funded software project that harnesses the power of collaborative development to fuel innovation across industries and ecosystems. https://www.dronecode.org
Comments
Rob
I am still trying to understand the options. I wish I had enough time to play with them but I still have a main gig that limits my time to do this.
Does Navio + not run a APM on Linux without a STM32? It does have a load of other things to deliver the I/O functions is a little pricey relative to the rock bottom prices of STM32 PX boards now. Seems a lot of people are gravitating towards ARM Linux and now we have Snapdragon Linux (which I am guessing is where 3DR is going).
And yes, EKF seems a bottomless sink of processor resources. I swear that my most reliable machine is an APM 2.6 with an LEA 6 running good old DCM.
Yep, or things like this:
http://www.inforcecomputing.com/6501-micro-system-on-module-som
@Marc here is a pixhawk and linux Co-processor
http://www.auvidea.eu/index.php/2015-11-08-08-01-27/can-centric/con...
@Jiro A module using a Snapdragon SOC might be possible for the modular motherboard from Auvidea
http://www.auvidea.eu/index.php/2015-11-08-08-01-27/can-centric/mod...
Marc, many of your comments are right on the mark. However, I will point out that the only code flying on the Intrinsync Qualcomm board right now, is in fact APM Flight Stack. Some are working on getting the PX4 Flight Stack working on it for some time, but the APM HAL layer makes porting to new hardware much easier. And don't forget that APM code has been flying on several Linux systems such as Raspberry Pi for some time.
I'll also highlight the fact that in order to fly the QFlight board, Tridge and Randy had to use an XRacer board as an interface between the ESC's and Qualcomm board. In this case, the XRacer is running a very simple sketch, simply doing UART <--> PWM translation. It would have actually been easier just to have APM:Copter running on the STM32 directly, but it was done this way, just to show that APM can run on the Qualcomm board.
The only case I see where an STM32 or similar chip would not be required to do IO translation to PWM, CAN, etc, would be in the case of a very simple quadcopter, with custom made ESC's controlled by some other communications method such as maybe I2C, which would be an extremely bad idea, BTW. Just ask any AlexMos/Basecam system user how much fun they've had dealing with I2C errors caused by signals running in proximity to power wires. And if you did this, there would not be any external IO such as Analog, etc, so the utility of the board for higher level research systems and hacking would be low.
The only time I see the STM32 not running the APM code directly, is if we get to the point of having 3 IMU's and 3+ concurrent EKF instances for higher levels of redundancy. That day will come. But again, when it does, APM is easy to port to any processor you want so... And I don't even see much point to that level of IMU redundancy anyway so long as GPS reception problems are far and away the most troublesome sensor errors we see.
Looks nice. I am sticking with my X2s and my VR uBrain for now.
@Rob. What you say is the conventional wisdom. Yet it seems 3DR is moving in the direction of Snapdragon FCs (if i read correctly between the lines). I am assuming that the Snapdragon FC that is flying now is using a Linux port of APM? There does seem to be a general acceptance of the all-in-one Linux box and the $5 Pi Zero will get that down to $70 or less when Banggood start selling variants of them. It does not seem that we will ever see Pixhawk 2 and Linux Co-Processor as no-one else seems to have followed Solo. I guess as UAVs morph into flying networked devices with more intelligence Intel and Qualcomm will be supplying the processing power as they do now with iPads and Androids.
And presumably QC will start to drive the development in Droncecode. Will we soon see a Solo with a QC FC? Will the STM32 APM Flight Stack continue beyond the current PX4 hardware setup (which seems to be being close enough to being reached).
Interesting times.
It's actually very easy to do. There are plenty of System On Modules available on the market, the one I'm using is only about $200. And then interfacing to a Pixhawk class hardware is trivial.
Keeping the higher level processor separate from the flight controller, just communicating via UART, means pretty much anybody can do it. I honestly don't know why anybody wants to complicate things by trying to do it all on one processor. Cell Phone processors are not designed to be robot controllers, and lack the IO necessary for things such as PWM, CANbus, etc. So you are required to add some sort of interface in any case. An STM32 processor is so small, I don't buy the argument that the systems are cheaper and smaller by omitting it. I believe I have the smallest system in existence, even including ESCs on a stack about 45mm x 100mm by 20mm. Maybe the one in the Beebop is smaller, I'm not sure.
Hi, Rob. That is great news as your involvement of such a interesting development.
It might be a huge development work other than the vehicle is just go well. I hope your experience will make some feedback to this community:-)
Jiro, I've helped develop a board comprised of a Snapdragon 805 processor used in a System On Module, added onto a carrier board containing an STM32 co-processor and IMU. It's actually quite a bit more smaller, and with native ability to do PWM input and output unlike the Intrinsync board. Unfortunately, there are no plans to produce it as a stand-alone product at this time, it will be used in a complete vehicle.
I am wondering, is there any person or group try to make board that is equivalent to Qualcomm Snapdragon flight board ?
A major omission here is that the APM flight stack is leading the way by already successfully flying on this board with current master!
Awesome job to the Ardupilot team!
Also missing, are the mentions of the other flight boards that Ardupilot runs on...
Pixhawk 1 and 2, VR-Brain, APM series, Pixracer, and many others.
Or the fact that the Linux port of APM began as the Pixhawk Fire project, the foundations of this allowed the process of moving to Linux on boards like the flight to be possible.
All done by Ardupilot team members and Friends.
I am personally now a member of DroneCode, on the Silver level. And am elected to the board of DroneCode to represent the members of the Silver level (Along with Jamie)
Both APM and PX4 are important players in this ecosystem, both fill different roles and both are equally represented by the goals of DroneCode.
-
1
-
2
of 2 Next