Week 2
07/10/2025
New week, new learnings, and more progress in the early stages of my Autonomous Driving (AD) studies. I previously ran CARLA Sim 0.9.15 on Docker using the ROS2 bridge with good results. I successfully scripted and ran simple examples for ego vehicle creation, autopilot, and driving without any issues. My GTX 1080 Ti has been supporting my work well so far!
This past week, I went back to the basics. I started CARLA Sim 0.9.15 directly from the terminal and operated the same examples listed above using scripts based on Python API, achieving identical results. This workflow is interesting due to some time-loss issues when working with the ROS2 bridge, as reported by teammates in the research group.
Intrigued by the fine-grained details of Sergio Paniego's PhD work, I took a deep look into papers [1], [2], and [3], where the contributions TrafficMetrics, DetectionMetrics, and BehaviorMetrics were presented. After reading these documents, I gained an overview of the models, architectures, frameworks, and datasets used.
[1] J. Fernández, J. M. Cañas, V. Fernández, and S. Paniego, "Robust real‐time traffic surveillance with deep learning," Computational Intelligence and Neuroscience, vol. 2021, Art. no. 4632353, 2021.
[2] S. Paniego, V. Sharma, and J. M. Cañas, "Open source assessment of deep learning visual object detection," Sensors, vol. 22, no. 12, p. 4575, 2022.
[3] S. Paniego, V. Sharma, and J. M. Cañas, "Behavior metrics: An open-source assessment tool for autonomous driving tasks," Sensors, vol. 22, no. 12, p. 4575, 2022.