FourPlusOne — Mini Autonomous Vehicle (OpenZeka MARC)
Dec 1, 2018
·
1 min read
Mini Autonomous Race Car (MARC) — Jetson TX2 + ROSProject Overview
The FourPlusOne team participated in the OpenZeka MARC competition, developing a mini autonomous driving system capable of navigating complex tracks. The system integrated lane tracking (CNN), traffic sign detection (YOLO), and LIDAR-based obstacle detection.
Key Features
- Deep Learning: CNN for lane tracking, YOLO for traffic sign recognition.
- Sensor Fusion: LIDAR + camera for robust perception.
- SLAM & Localization: ROS-based mapping and navigation pipeline.
- Edge Deployment: Optimized on NVIDIA Jetson TX2 with TensorFlow, OpenCV, and ROS.
This project demonstrated the potential of autonomous driving systems in constrained embedded platforms, and received recognition in the MARC competition.
