Home Creators Posts Import Register Favorites Logout
haven't archived this post yet. have a subscription? use the importer!

Content

This is the current state of openDog - I've made cages that each hold three ODrives, and these fit neatly inside the body of the dog. The electronics sit on top for now, mainly so I can get to them for development, although there's actually space for the microcontroller elsewhere ;-)  Quote a lot of wiring to do still.

I'm using Teensy 4.1 this time due to it's large amount of serial ports:  https://www.pjrc.com/store/teensy41.html 

I'll be looking at Nvidia Jetson Nano in parallel, but that will probably perform some high level functions eventually and be used in other builds too.

Files

Previews only

Comments

Neil Devonshire

That's a very nice looking processor, it dwarfs my Mega 2560 and has plenty of serial ports, plus very cheap. Have you got the additional memory chips on the bottom? Would love to know if you are using the memory in this project. He's coming along nicely, have fun with the wiring.

Eric lamb

What are your thoughts about using CAN bus? I have a firmware for the Odrives 3.6 56v if you need it. That way you don't have to compile it on your own. (that's a pain)

Simon

are the odrives still not able to do CAN?

XRobots

No it's stock at the moment, yep - it's pretty much the fastest Micro you can buy right now

XRobots

I'm reusing the 48V ODrive 3.5s from openDog V1 right now. I'm just going for serial again for now.

XRobots

It's working in Dev but it's not in production yet. I'll wait until there's an Arduino Library and someone else has worked out the bugs etc.

Eric lamb

I also have the 48V ODrive 3.5 CAN bus firmware and I can show you how I got It working if you change your mind, But if you are happy with using serial then I wouldn't bother changing anything.

Anonymous

I guess 2D data from LIDAR can help with obstacle avoidance. The Jetson Nano would be super useful for visual recognition!

Anonymous

I've used that on an autonomous rover before and my experience is that 2D Lidar only works well on flat ground, as soon as you have slopes or uneven ground you want 3D imaging. Also I can imagine that a walking robot very much amplifies this problem as you have to plan things like where you want to put your feet. But hey, 2D Lidar object avoidance is better than no object avoidance, so why not?

XRobots

I'm just going to leave it as it is for now - that is tried and tested and I know exactly how to make it work ;-)

XRobots

Haven't decided exactly what to do yet, it'll be longer term for this project