[r2_pilot]

I have that exact problem right now. I've got a bespoke 16 pound robot chassis with Jetson Orin nano and a Teensy, and there's all kinds of sensors, and it's connected to the motor controller with wheels, a DepthAI camera that can identify individuals, and at the end of the day, it's just a bunch of loosely coupled components lol at least I'm almost done with the main integrations and can focus on the way the subcomponents interact.

[taneq]

Oh no, soon you’ll have to actually make it so something fun or useful, which is a day job for a team of engineers and even then they usually make a hash of it.

I have a disassembled BigTrack, an esp23-cam board, and a bunch of motor controllers in a box. I’m planning to make it follow lines and do things when you show it QR flash cards. Same kinda thing but it takes up less room. :D

[r2_pilot]

I originally started this shortly after(before?) Google showed their multimodal Palm model or whatever; while I have some experience with ROS2 and will be doing most of the fun (it is for fun-- as it stands I've learned a lot about how to respect this level of equipment) I am planning to let AI test drive it at some point.

[taneq]

Oh, please don't take my comment as questioning your actual ability to do real world useful stuff with it! :) I just meant there's a definite inflection point where it goes from "whee I'm having fun building a robot" to "oh my god did this just become my day job" and from personal experience it's really easy to lose steam at this point. :P

What are your plans for this beefy rover?

[r2_pilot]

It's more of a mast-bot than a rover! [edit: I did hit the point where it got to be more work than enjoyment; that's when I upgraded the Jetson Nano to the Orin line and it became fun to work on again ] Think a 2wheel rectangle (those motors were from a defunct telepresence bot whose proprietary batteries failed and too expensive to replace; this kickstarted the whole project), with 2x 1 meter 2020 aluminum on the front/rear midpoints of the rectangle. I just wanted a general purpose platform, but then my work did a creative staff residency in which my proposal to make this bot play my trumpet was accepted; this has caused a lot of design changes (while enhancing the original platform where feasible). I've gotten the basics(proof of concept, all individual pieces tested) of that completed and now it's just a matter of integration, again and again amen lol. Technically I can drop that trumpet playing module as I've spent my time budget allotted for that particular day job crossover, but as far as it's come and with as "little" work to go, I'm probably gonna finish it up in the next weekend or two. Then back to the AI stuff. It'd be cool to do a sock sorter (bane of my existence right there), but I'd also like to get it to do other light household chores. In addition, I'm using it to test the Jetson Orin product line for offline AI suitability. {if you've made it this far-an original goal from before this particular bot is to have it identify when people are standing in a circle and provide party favors, while maintaining a circularly linked list if the circle dispersed, without going into more details =] }

[niemandhier]

If it has the needed actuators: Teach it to sort and match socks, that's boring and somehow needs to be done all the time. My life would be better if I had a sock sorting robot.