Autonomous Experimental Rover · v0001

T1 SKYNET

An experimental edge-AI robot — it sees, maps, navigates and acts on its own hardware, out in the world. Perception, navigation and control, running on the edge.

Access is by invitation only · granted by the operator

The story

Built by hand in 2018.
Now written entirely by AI.

T1 started in 2018 as a personal project — a four-wheeled rover built to experiment with robotics and artificial intelligence, and to learn how perception, navigation and control come together on real hardware.

Today, T1 is programmed entirely by large language models. Its software — from the ROS 2 stack that drives the robot to the interface you're reading now — is written, extended and maintained by AI such as Claude. The human sets the direction; the AI builds the robot.

What it is

An edge-AI robotics platform

Core platform

Raspberry Pi 5 with an AI accelerator HAT, running ROS 2 robotics middleware.

Edge AI

A 26-TOPS neural processor runs on-board object detection and face recognition in real time — no cloud round-trip.

Vision & sensing

Intel RealSense D435i depth and T265 tracking cameras feed a forward multi-sensor fusion array.

Control & mobility

A real-time microcontroller handles low-latency motion on a four-wheel rover chassis.

Connectivity

4G LTE / Wi-Fi keeps T1 online, so authorized operators can watch and command it from anywhere.

Mission

A test bed for autonomy experimentation — edge AI, perception, navigation and control.

Raspberry Pi 5·AI HAT · 26 TOPS·ROS 2· RealSense D435i + T265·4-wheel rover·Edge compute · vision · control

Watch T1 think.

See what the robot sees — live camera with AI object and face detection, position, and depth — and give it tasks by voice or text.