Driven by vision. Defined by persistence.
B.S. in Automotive EngineeringKookmin University, Senior
The physical foundation of the autonomous driving stack — vehicle platforms, sensor suites, and onboard compute units that interface with every software layer above.
Two primary platforms — ERP-42 (full-scale autonomous vehicle with CAN-based steering and throttle control) and JetRacer (1/10-scale RC platform for rapid RL prototyping and Sim-to-Real transfer experiments). Both share a common ROS2-based software interface.
Multi-modal perception hardware including Velodyne VLP-16 LiDAR for 3D point cloud generation, monocular USB cameras for lane and object detection, GPS/INS modules for global positioning, IMU for inertial measurement, and wheel encoders for odometry feedback.
NVIDIA Jetson Orin serves as the primary onboard compute unit, running ROS2 nodes for real-time perception and control. Host PC handles heavier tasks like model training and simulation. Communication between onboard and host uses a high-throughput bridge layer.