| Project Domains | Mentors |
|---|---|
| Robotics, Autonomous Driving, Imitation Learning, Embedded Systems | Vrushtee Gaikwad, Sahil Apage |
Project Description
This project’s goal is taking an off the shelf RC car and building it into a self-driving vehicle that learns to drive from a single camera bringing the end to end approach proven on full size cars down to a small vehicle whose entire control stack we build and own inspired by modern full self-driving (FSD) paradigms.
The system relies on an onboard edge computer acting as a centralized electronic control unit (ECU) that processes raw camera feeds through a deep neural network to directly output control commands to steering and throttle, trained on demonstrations we record by driving the car manually and over remote teleoperation.
To fit a strong policy on small silicon, a large model trained off board is distilled into a small student network that runs in real time on the car.
The final output is a single camera RC car that drives a real route autonomously, accepts remote teleop takeover — a self-built autonomy ECU that could drop into any small vehicle.
Resources
Learning to Drive from a World Model
simcity
openpilot
commavq
ecu
The first robotics agent fully trained in a learned simulation