I am tempted to buy one and try it. The bang for the buck ratio seems to be attractive....
However
"UNO Q --> Game changer" Well, that's what their marketing is saying.
My experience with Linux based boards is not good however. You see some really impressing project and think you can easily modify it to your idea.
But then I normally work with Android and Windows and have to spend time for Linux alone. Then it turns out that that project does not work anymore, because some well meaning idiot -sorry, meant enthusiast- "improved" something deep (for me) in Linux and you can't compile anymore. Then it turns out, that the documentation of the library is just not good enough and after you have managed to understand, you find the library does not cover your need. And you are stuck with ultra slow Python. Often things work "to some extend".
I think, that many professionals work with Linux and therefore can't see that this alone costs a lot of time for others. And of course the shining desktop is tempting.
UNO Q might have some integrated IDE but still is using three totally different worlds: Arduino, Python (or other) and some AI tool. I can't imagine, that there is time in education classes to learn these 4 (if you count Linux too) worlds.
What I find strange with Uno Q is, that they did not bring out more ports of the microcontroller. Some pins are unavailable and many are used for this LED matrix.
Found this project: https://www.hackster.io/naveenbskumar/ball-chasing-quadruped-robot-with-ros-2-arduino-uno-q-cf9701 Well, they use a ball with a very distinct colour there and have to hold it directly under the "dog's" nose. That project is done by a professional researcher and he uses 2 additional ESP32 boards. I do think, that you can achieve the same and better with your new SPI camera and P2 if you use colour blob detection and forget AI! The strange thing is, that "we use AI" alone makes something good these days....
They are saying that this is all about education. But this is all so extremely ineffective. Python burning power, AI burning power, using 4 Linux cores + 2xESP32 + 1xM33. Is this really what we want to teach the next generation of engineers?
Comments
I am tempted to buy one and try it. The bang for the buck ratio seems to be attractive....
However
"UNO Q --> Game changer" Well, that's what their marketing is saying.
My experience with Linux based boards is not good however. You see some really impressing project and think you can easily modify it to your idea.
But then I normally work with Android and Windows and have to spend time for Linux alone. Then it turns out that that project does not work anymore, because some well meaning idiot -sorry, meant enthusiast- "improved" something deep (for me) in Linux and you can't compile anymore. Then it turns out, that the documentation of the library is just not good enough and after you have managed to understand, you find the library does not cover your need. And you are stuck with ultra slow Python. Often things work "to some extend".
I think, that many professionals work with Linux and therefore can't see that this alone costs a lot of time for others. And of course the shining desktop is tempting.
UNO Q might have some integrated IDE but still is using three totally different worlds: Arduino, Python (or other) and some AI tool. I can't imagine, that there is time in education classes to learn these 4 (if you count Linux too) worlds.
What I find strange with Uno Q is, that they did not bring out more ports of the microcontroller. Some pins are unavailable and many are used for this LED matrix.
Found this project: https://www.hackster.io/naveenbskumar/ball-chasing-quadruped-robot-with-ros-2-arduino-uno-q-cf9701 Well, they use a ball with a very distinct colour there and have to hold it directly under the "dog's" nose. That project is done by a professional researcher and he uses 2 additional ESP32 boards. I do think, that you can achieve the same and better with your new SPI camera and P2 if you use colour blob detection and forget AI! The strange thing is, that "we use AI" alone makes something good these days....
They are saying that this is all about education. But this is all so extremely ineffective. Python burning power, AI burning power, using 4 Linux cores + 2xESP32 + 1xM33. Is this really what we want to teach the next generation of engineers?
Sorry, Christof
I think it’s all about the TOPS…
Sounds like it is unmatched at this price.
The Linux performance is not as good as rpi. But that’s not what one might care about if interested more in AI performance…
Ok, I’ll change game changer to a question …
Thought it had an NPU but it doesn’t .
But, can use GPU for tinyML things so that is something close…
I heard a tops rating but now can’t find it…
If that example needs more MCUs to work, might be opportunity for P2. Have to look at that…
Looks like Arduino VENTUNO Q is coming soon. This one does have NPU with TOPs rating.
Was looking to compare with Pixy2, this looks to be much stronger than that.
But, came across Huskylens2. This does have a TOPS rating of 6..10 looks like.
This might be better fit for P1/P2 robot...
The cost is a bit more, but comes with camera and display....
Huskylens can connect to Gravity $16 voice recognition module. This might be more interesting...
Also found a low cost Gravity face/gesture recognition module...
Things things might be better suited for P1/P2 robot and are easier to use...
Ok Parallax already sells huskylens.. I have to catch up on this stuff…