..
RC Car Multi-Agent RL
A decentralized architecture for coordinating a team of 1:64 scale RC vehicles on cooperative tasks, built for real-world reinforcement learning training rather than pure simulation.
- ESP32 microcontrollers handle inter-agent communication, tuned for reliable, low-latency data transfer since coordination breaks down fast once messages lag.
- Control policies run per-agent, coordinating the fleet toward cooperative objectives.
- A Python-based simulation environment was used to implement and validate the multi-agent RL framework before deploying to hardware.