How teketesfaye/speed-control-based-on-detection-result works
Project Type and Purpose
• This is a ROS2‐based C++ package (“speed_controller_cpp”) that implements an autonomous safety system.
• It subscribes to object detections from one or two YOLOX vision nodes, merges/fuses detections, analyzes whether any object lies in the planned trajectory, then publishes speed commands (including emergency stop) to the vehicle’s controller via Autoware.
File Structure
Top‐level:
– scripts/ (setup script)
– speed_controller_cpp/ (ROS2 package)
Inside speed_controller_cpp:
– CMakeLists.txt, package.xml (ROS2 build & metadata)
– include/speed_controller_cpp/ (public headers: speed_controller, core logic)
– src/ (node entrypoint + core implementation)
– config/stop_line.param.yaml (tunable behavior parameters)
– launch/speed_controller.launch.py (launch file defining ROS nodes & parameters)
README for Additional Insights
• Dual‐camera safety system using YOLOX for object detection
• Real‐time detection fusion via IoU overlap logic
• Publishes ROS2 topics: detection inputs, fused detection outputs, speed_command outputs
• Integrates with Autoware trajectory planning
• Architecture line: Camera(s) → YOLOX Detection → Object Analysis → Speed Control → Autoware Trajectory