Wow! All I can start off with about Snickerdoodle is BAM! In your face computing power!
This is what the next generation of DIY Drones and Robotics looks like! All I can say is AWESOME!
This product is so new and so innovative it's unbelievable!
Guy's I've been doing research lately looking for hardware for a cubesat project (small satellites in low earth orbit) and trying to figure out what combination would best serve that purpose but this product is excellent for drones!
That's not all!
I've found some open source software at NASA!
Here we go again with the BAM! Check this out! Open source middleware from NASA for flight control!
Which works on top of!
RTEMS!
RTEMS is FREEWARE which was developed for U.S. Army Missile Command and RTEMS initially stood for Real-Time Executive for Missile Systems, then became Real-Time Executive for Military Systems before changing to its current meaning. How cool is that?
An inexpensive and innovative OPEN solution with the "Right Stuff" to take drones and robotics to the next step!
Here we have a triple threat!
Something drones need to progress, more room to develop the applications you want! And the power to carry them out! In a very small package! This combination is pure dynamite for a progressive drone with CERTIFIED software that will get drones what they want politically as well as provide real stability and safety based on certified and tested software.
One more thing I might add is there is a protocol specification SPA ( Satellite Plugnplay Architecture - AIAA S-133.)
It's an open standard, Basically, each component (IMU, Radio, RCS, etc) has a small processor
on board that 'knows itself' with its capabilities and calibrations.While many use ARMs and stuff, there were some examples of AVRs.
If you add that to the mix all external hardware becomes plug n play and as you grow you can add hardware and there is no need to configure it. The little processor on the hardware negotiates with the software and self configures.
GO GET IT! :)
Relevant links:
http://krtkl.com/ (pronounced critical)
Comments
No, but I will keep trying until I get a hit! Plenty of things to try! I did finish the 5 Axis CNC project! But it's in use at the moment on a private project. hopefully I can release that soon. I just started another project to open source the Chevy Volt!
Hi, did you get any progress in this project? Thanks
Resistance is futile you will be assimilated. lol
Hey! Guess what! I got a little team together and we are working on this.
Contact monroe@teamprometheus.org if your serious about wanting to see this happen.
It is inevitable you know.
FPGA is coming.
GET YOU SOME! :)
But we are starting with Beaglebone RTEMS and CFS first. PX4 devs are interested one joined up.
The Parallella could be an option as it is available now. Let's see how Sinkerdoodle does!
Cutting & pasting Arduino code is fun and easy , Python on Linux is the next natural step if you are looking to build fast prototypes. But Verilog on FPGA is like the big leagues, no more pee wee stuff, you really have to get an intrinsic knowledge of both hardware and software and opensystem ressources are mostly limited to a few forums and the genius work of a hamster (this one for you Mike Field).
If you plan to get an experimental RTOS on top of that and use a decmmisionned NASA middleware that was probably a good framework in the pre y2k era, its like drinking too much Xmass Punch, you can get pretty high, but it might be a hard awakening.... By the way, were is Monroe?!?
Make no mistake. I also get the enthusiasm. But the ever growing pile of dev-kits for cool hardware that I never have time to play with, have thought me some hard lessons about time and hardware vs. software.
For example, my personal favorite would still be the Parallella board. https://www.parallella.org/board/
It's $99, and has the same Zynq FPGA with dual core ARM A9 as the Snickerdoodle, but in addition there is also a 1ghz 16-core Epiphany DSP chip.
Agreed this sounds pretty awesome. I get your enthusiasm Monroe King...
@John @Patrick, we're on the same page with this. We see FPGAs being used in specialized hardware like smart sensors and smart actuators where very high reliability but limited functionality is needed. We've been using them for years and I absolutely agree that the long development time and limited availability of experienced developers is a real problem for more generalized and flexible use cases. IMHO, systems like the Snicker' make the barriers to entry higher without really adding novel functionality when compared with the secondary computer route.
Patrick is touching on an important point. Strangely enough, new fancy hardware is dime a dozen these days.
But robust and proven autopilot software is not. And the threshold to make such software is higher then most other software project, limiting the effectiveness of community driven development.
With this in mind a FPGA would be dead in the water. While being much simpler then it used to be, there is still to much specialized knowledge and software required to deal with them.
This is why most devs are going the Linux + independent sensor board (aka secondary computer) route. It's best of both worlds, since you get to reuse the already proven APM code for flight critical stuff and have a generic Linux environment that more people are familiar with on top for high level applications and the next gen experimental stuff.
It would be a dream to have such level of standardization across software and hardware, including IMUs & sensors. But my opinion is that we'll never get there because of always the same reason with humanity: money. Who's interest it is to standardize ? only naive utopists like me, not industries.
Apart from that, I find quite old tech aspects in this approach (they kickstarted two years ago and it shows...) : a dual core ARM processor not even running at 1Ghz is obsolete, a SPA standard looking like an eighties technology (this is volontary in space programs to run stuff that date back 10 years earlier, to know precisely how things run. They never use the latest technology as no reliability/deep behaviour knowledge exist for latest new tech; question of reliability).
-
1
-
2
of 2 Next