Run YOLO Object Detection on NVIDIA Jetson with TensorRT and ROS 2

These tutorials guide users in creating ROS 2 perception pipelines on NVIDIA Jetson. One approach utilizes YOLO with TensorRT, emphasizing the separation of camera capture, inference, and robot decisions for independent optimization. Another method employs DeepStream for multi-camera AI object detection, focusing on reducing CPU copies by keeping decoding and tracking within the GPU-accelerated pipeline. Both paths provide a development workspace, a testable robotics pipeline, and integration options for Flutter, including logging and troubleshooting practices.