JetBot, a $250 DIY Autonomous Robot Based on Jetson Nano Impresses at GTC

by Brian Caulfield

Even at a conference packed with sophisticated autonomous machines that walk, drive, fly and even slither, on their own, the $250 JetBot was a standout.

Based on the Jetson Nano, the small but mighty $99 AI computer introduced by NVIDIA CEO Jensen Huang at GTC last week, the JetBot drew a crowd of hundreds to a session where its creators explained how to build one of your own.

The bill of materials? Just $250, including the Jetson Nano. That includes a camera, motor and motor driver, and even a tiny PiOLED display.

Yet the dinky robot is capable. The Jetson Nano powering it supports high-resolution sensors, can process many sensors in parallel, and can even run modern neural networks on each sensor stream — giving the JetBot some amazing capabilities.

“With JetBot, you learn not only the training and deployment of deep learning models, but also how to collect a dataset,” said Chitoko Yato, the JetBot’s co-creator. “We run through the full workflow for teaching the robot to avoid collisions by labeling images captured using the onboard camera.”

Bot to You by Jetson Nano

The Jetson Nano that the JetBot is built around comes with out-of-the box support for full desktop Linux and is compatible with many popular peripherals and accessories. Its ready-to-use projects and tutorials help makers get started with AI fast. The small but powerful CUDA-X AI computer delivers 472 GFLOPS of compute performance. Yet it’s power efficient, consuming as little as 5 watts.

All the instructions to build the robot with Jetson Nano are shared on GitHub, so it’s easy to get started. Once you do, you’ll be able to enjoy education tutorials from basic motion to AI-based collision avoidance. And you can interactively control it all from your web browser.

At GTC, John Welsh, a JetBot co-creator, showed it off to hundreds of gawkers as it wound its way through a miniature Lego city.

“It’s all open source, the hardware, the software,” Welsh said. “Then you can take what you learned, take the components and you could build something new.”

Who knows where JetBot will take you.