Which microcontroller/computer to use? RPi, Arduino, which model #59
Replies: 3 comments
-
Someone suggested that I reconsider using RPi 4B and suggested that the situation has no signs of getting better in the near future. |
Beta Was this translation helpful? Give feedback.
-
The discussion with Helen Leigh from Crowd Supply:
That's a good point about the RPi 4. I think I'll try to build the system around a Pico W with ML training/decision-making happening externally. Though, there is definitely the appeal of having some reasonable compute power to run more expensive ML algorithms on the RPi 4 for decision-making (i.e. make it a truly standalone system). Open to suggestions for other single-board computers than the RPi 4 if you have any (still with "reasonable" compute power). Of course, I can look at articles like https://www.zdnet.com/article/best-raspberry-pi-alternative/, but it's tough to know what might be a better alternative for this specific project where ease-of-use is a high priority.
|
Beta Was this translation helpful? Give feedback.
-
Getting Blinkt! to work with the Pico seems like a pain. See this discussion. There's a guide for controlling an LED strip with a Pico. For the Pico, I could use a terminal block breakout module, or maybe even better the Maker Pi Pico Base since it has a single NeoPixel LED and the Grove ports can connect to QT sensors via a Grove->Stemma-QT adapter. There are a couple of outstanding issues with switching to a Pico design:
I looked into Banana Pi M2 Zero, and it seems a lot less fluid getting it set up. See for example the NeoPixel RGB guide. I also emailed RPi asking about fulfilling a Crowd Supply campaign. Crowd Supply requires a final(ish) fully-functional prototype, and so I'll probably need to get my hands on whatever RPi I want to be the final design for the campaign. Subscribing to rpilocator Twitter feed to see if I can get a Pi Zero 2 W or Pi Zero W. |
Beta Was this translation helpful? Give feedback.
-
https://twitter.com/JoshuaSchrier/status/1543374043230904321
Joshua Schrier suggested Pico, offload the decision-making to a computer. I was thinking it would be nice to be able to run the module standalone, with it connected to WiFi. Probably not bad to have a USB connection for the demo, but I wonder if it might be better to have it connected to WiFi for the scale-up/transition to a "real" task. Especially if there are going to be interactions/submissions via Google Colab. In this case, a free Colab session would probably time out by the time the real experiment has completed the optimization.
Beta Was this translation helpful? Give feedback.
All reactions