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

Content

Next week's video for Patrons is about adding a Jetson Nano to my tracked robot. It's a really basic demo doing the most basic 'transfer learning' possible to retrain a deep learning model to look for specific markers. I'm just getting my feet wet with the tools and scripts provided by NVIDIA before I do anything more complicated - there's plenty more projects coming that'll use machine vision with deep learning models.

The process was surprisingly easy and works pretty well - video comes up early next week for Patrons.

Files

Previews only

Comments

Anonymous

I see you changed out the motors. Those are surprisingly torque-y

Anonymous

Hi James. Can you provide a list of the parts for the build? I’d be really helpful as I’m building a robot myself for a university project and i’ve got zero experience in this field. Keep up the great content. Kind regards, Mark

Anonymous

Hi James, I highly recommend getting a usb 3.0 ssd drive to plug into the jetson nano and use that for your work. I have found that doing heavy write I/O operations to the standard micro SD card can really cause the jetson nano to bog down and lock up while the write is occurring. Isolating your work and disk I/O onto a separate usb drive will (I think) save you some head aches... Good luck!

XRobots

Thanks, it seemed smooth enough for this, but I'll consider it. I believe the Xavier NX has a slot for an SSD which I may use.

XRobots

The CAD and Code will be up on Git shortly, I'll try to document the other parts - there's not a lot to it, just some BTS7960 drivers and some motors.