CARLA Base Controller
GitHub: Highsky7/CARLA_Base_Controller Context: B.S. Graduation Research Project
Overview
CARLA Base Controller is a vehicle autonomy framework built on top of the CARLA open-source simulator with a ROS1 bridge. The project focuses on establishing a reliable closed-loop vehicle control baseline that can be extended with higher-level perception and planning modules.
Key Contributions
Integration — CARLA + ROS1 Bridge
- Set up the CARLA–ROS1 bridge to synchronize simulation state (sensors, odometry, control commands) with the ROS topic/service ecosystem
- Enabled modular development: perception, planning, and control nodes communicate via standard ROS messages
Longitudinal Control — PID Controller
- Implemented a PID controller for throttle and braking commands
- Tuned gains for smooth speed tracking across varying road gradients and curvatures
Lateral Control — Pure Pursuit
- Designed a Pure Pursuit path-tracking controller with adaptive look-ahead distance
- Validated lane-keeping performance on CARLA’s built-in map scenarios
Technical Stack
| Category | Technologies |
|---|---|
| Simulator | CARLA |
| Middleware | ROS1 (CARLA–ROS bridge) |
| Longitudinal Control | PID |
| Lateral Control | Pure Pursuit |
| Language | Python |
Outcomes
- Established a reusable autonomous vehicle control baseline in CARLA
- Demonstrated stable closed-loop driving on standard CARLA town maps
- Served as the foundation for further research into perception-integrated autonomy