Some non-technical reasoning I believe in strong decouplig of the autopilot and the companion (this doesn't mean they can't live within the same box). However plug&play is a great plus, as this means that I (final user) can buy an Iris / Pixhawk / whatever and in a second moment I can add the companion as a second layer of computation, while all the underlying settings remain the same. This mantains a very attractive learning curve. In a third moment I can even switch to a more powerful companion as new needs / new supercomputers arrives.
In this group we've been dealing with companions fore a long time. Now @Randy opened a repo called "Companion" https://github.com/diydrones/companion. A very big thank you for this! I invite everyone to contribute.
SUGGESTION ABOUT COMPANION REPO:
- COMMON (this is a directory with common scripts. @Randy maybe also a simplified version of balloon fineder could fit here: colour_finder_web.py ). I'll put also VPN scripts, gstreamer (@Bernt please contribute!!) I can also contribute with 4Gmetry .py scripts.
- UBUNTU / ODROID XU4 this is what we're doing here at 4Gmetry until now.
- UBUNTU / JETSON TK1 this is the supercomputer. As the main usage will be computer vision, I think that we should give a look at http://elinux.org/Jetson/Computer_Vision_Performance as suggested by @Bill_Bonney
- n OS / BOARD combinations
- PERFORMANCE this includes performance tests across the above combinations.
What do you think?
Replies
I like the decoupling of boards as well- our group has had a lot of success with both apm/pixhawks talking with separate onboard computers. Also from safety side I think it makes more sense. Other computers that should be looked at are Gigabyte Brix's. They cost on order of the new Nvidia and they are Intel with beefy cpu. We use them and love them. We have used arm based cpu's in the past and not having the ability of using the same setups as you do on the desktops makes life a bit more annoying. The downside is weight but they approx 200g. Anyway something to consider if you need the additional cpu
I just wanted to mention that I have received great service from the Volta team. Thanks a lot Silvio
Silvio just started two threads for TX1 and TK1 discussions. These should be sufficient for near-term development. If the TX1 development proves to be fruitful, it may eventually deserve its own "user group".
TK1 Discussion: http://bit.ly/1MIVsaw
TX1 Discussion: http://bit.ly/1kwkrDK
Miniature TX1 carrier board designs proposed by Jurgen from Auvidea. Details are over here: http://bit.ly/1MOBKWx
Your feedback would be appreciated.
Conversation is going to explode. :D
As someone suggested, I splitted for the 3 main topics/boards:
XU4 http://diydrones.com/group/volta/forum/topics/companion-xu4
TK1 http://diydrones.com/group/volta/forum/topics/companion-tk1
TX1 http://diydrones.com/group/volta/forum/topics/companion-tx1
Comparisons stay here.
Gitter Channel https://gitter.im/silviorevelli/companion
On this video, showing all the promising products based on NVIDIA, you can see the Percepto board.
Here what Percepto wrote on Indiegogo recently:
...we identified a need to separate the user control interface from the processor to create a separate drone control path that is less sensitive to software issues. Practically this means that we have added an FPGA to our hardware! we use this FPGA to handle all tasks that require uncompromisable robustness like flight controller inputs......
That is a reminder that the best architecture for an autopilot shoud be a balance between the very adaptative programmable logic of FPGA and brute processing power of GPUs (and this is not taking in account the talks about all the KS startups that are facing the fact that DRONE IS HARD and having to change the architecture of a system on the fly is quite a challenge....best of luck)
Patrick
Thanks. I have definitely followed the Percepto campaign. I am sad that it was delayed, but a few months is not so bad.... I am a little confused about this issue. If a reliable autopilot and companion computer are separate modules (and the user interfaces directly with the autopilot), does that address the issue that is being raised? I am not trying to be negative; I honestly am looking for some clarity on this point.
Also, I got in touch with Colorado Engineering. They would like some feedback from us. I will shoot some more questions your way in the near future.
Good move for Colorado Engineering, thanks
Thanks. I *think* I understand .... I envision that my next year of development will primarily be on a system such as this: Pixhawk2 w/ dedicated radio link, directly connected to TX1 (USB -> FTDI -> Serial), w/ new development peripherals connected to TX1.
Also, I am interested in whatever Qualcomm is coming out with, but details seem to be limited at this point.
I will be on the phone with Colorado Engineering on Monday. If you have any particular questions, let me know.... It seems like they are an extremely capable group. I am sure they could meet virtually any of your requirements.
-
1
-
2
-
3
-
4
of 4 Next