Project SCOUT Explore the project
+ A DIY project. With AI in the workshop.

Big ideas.
Real wheels.

White Project SCOUT ride-on Jeep with chunky tires, black roll cage, rooftop LiDAR and a front camera
A familiar ride-on.
A new reason to learn.

A DIY family rover.
Built by learning,
with a little help from AI.

See the idea in motion
SMART COMPANION FOR
OUTDOOR UTILITY & TODDLER TRAVELDIY robotics · Learning with AI
+ MADE TO LEARN
BUILT TO SHARE

Some ideas deserve
to leave the screen.

SCOUT is a DIY hobby project turning a ride-on Jeep into an outdoor companion rover. It explores what you can learn with AI as a research partner, coding helper, and patient explainer. The aim: take a question at the keyboard and turn it into something you can build, understand, and enjoy with family.

01 FROM IDEA TO MOVEMENT

The idea, in motion.

The ambition: a little companion for family walks and backyard exploring. Try the simulated Follow and Roam modes to see how sensing could guide the way.

THE BACKYARD / ROAM DEMO
Planned route SCRIPTED 3D SCENE
WHAT THE SENSORS SEEFollowing the backyard route.Blue dots mark simulated LiDAR returns

A little room to explore. Roam illustrates a route around the planter, with space to turn and a boundary to stay within.

Switch on a sensor. See what it adds.
LiDAR / DISTANCE IN EVERY DIRECTION

A way to measure the world.

A rotating LiDAR samples distances all around the Jeep. Blue dots show where the simulated scan meets a surface. The learning challenge: turn those measurements into a useful picture of the space ahead.

CAMERA / IMAGES INTO INFORMATION

From pixels to a path.

The front camera gives the computer images to work with. Following a person means finding them in those images, estimating their direction, and combining that with the LiDAR view before choosing a route.

GROUND / A CLOSER LOOK BELOW

The ground matters, too.

Seeing ahead is only part of the problem. Downward sensors check the surface near the wheels; bumper inputs add contact information. Both belong in the layer that decides whether motion should continue.

02 LEARNING UNDER THE HOOD

Real parts.
Real learning.

A Raspberry Pi, an ESP32, a handful of sensors, and the original ride-on hardware. Each part opens a door into power, code, or mechanics. Pull them apart to see how they fit.

SCOUT / COMPONENT STUDY
DRAG HORIZONTALLY TO ROTATE
0%
ASSEMBLEDEXPLODED
03 WHERE CODE MEETS HARDWARE

How an idea
becomes motion.

A robot needs energy, information, and a way to act on both. Follow the system concept one connection at a time—from a sensor reading to a turn of the wheels.

CONCEPT ARCHITECTURE
PowerDataControlStopping
THE CONTROL LOOP
01 / 20FOLLOW A CONNECTION
SELECT A COMPONENT TO TRACE ITS CONNECTIONS

Request. Move. Measure.
Learn from the difference.

Feedback lets the rover compare requested motion with what actually happens. Building it is a learning loop, too: ask questions, use AI to explore an approach, test it in the real world, and bring what you discover back to the workbench.