Racer is a complete open-source hardware and firmware project for building a small autonomous car based on an ESP32 microcontroller.
The project addresses the challenge of creating a programmable miniature vehicle by providing all necessary components: firmware written in C for the ESP32-H2, PCB schematics and gerber files for manufacturing, 3D-printable mechanical parts for the enclosure and wheels, and Python scripts for training neural networks to control the car autonomously. The car can be driven via a physical remote controller, keyboard input, or trained neural network models that learn from collected sensor data.
This project suits makers and hobbyists interested in robotics, embedded systems, or machine learning applications. It provides a complete bill of materials and manufacturing specifications, making it feasible for someone to build a working unit from scratch. The project includes both manual control options and an extensible framework for autonomous driving through neural network training. Someone considering this should be aware that it requires familiarity with hardware assembly, ESP32 development, and potentially Python for the machine learning components. The mechanical parts benefit from access to both FDM and SLA 3D printers for optimal results.
Development activity shows the maintainer actively engaging with the community through video documentation and responsive communication about the project's evolution. The project structure deliberately avoids git submodules despite spanning firmware, hardware, mechanical design, and scripts, keeping everything in a single repository for accessibility. The maintainer has indicated areas for future contribution including cleanup, color-sensor-based line-following functionality, and integration with large language models for robot control.
