← All projects

Open-source infrastructure

CARLA–Apollo Bridge

A bidirectional interface connecting simulated sensors and vehicle control to the Apollo stack.

Open source · Simulation and system integration

  • CARLA
  • Apollo
  • Docker
  • C++
  • Python
GitHub repository ↗Watch demo ↗

CARLA–Apollo integration demonstrations

Ground-truth obstacles

CARLA simulation alongside Apollo visualization of ground-truth obstacles and planned paths.
Ground-truth obstacle visualization in Apollo alongside the CARLA simulation. Open original animation ↗

Stopping at a red light

CARLA and Apollo views of a red-light stopping scenario.
A red-light stopping scenario shown in the connected CARLA and Apollo environments. Open original animation ↗

Signalized T-intersection

Apollo route and traffic-signal visualization with a CARLA view of a T-intersection.
A signalized T-intersection scenario with the simulator and Apollo views displayed together. Open original animation ↗

The problem

Testing autonomy software requires realistic sensor inputs and a reliable path for control commands to reach the simulated vehicle.

My contribution

  • Built a bidirectional interface for sensor-data exchange and control integration.
  • Tuned integration for vehicle dynamics and closed-loop scenario testing.
  • Supported workflows spanning simulation, hardware-in-the-loop testing, and deployment.

Technical approach

Sensor data travels from CARLA to Apollo, while control commands return to the simulated vehicle. This connection enables repeatable testing of the driving stack.

Results & validation

Closed-loop simulation and validation

The open-source bridge supports reproducible scenario testing and has been used by the research community.

Sensor/control architecture diagramSupporting material coming soon