Thursday, May 2025

03:30 PM - 03:50 PM

Room: LL20A

Session: Artificial Intelligence for Automotive Displays and HMI Technologies

Real-Time ADAS Visualization Using DL-GSA-Based Computer-Generated Holography

Student

• Chien Yu Chen

  National Taiwan University of Science And Technology

Description:

This study investigates the integration of Computer Generated Holography with Advanced Driver Assistance Systems for applications in augmented reality head-up displays. By utilizing the Deep-Learning Gerchberg-Saxton Algorithm (DL-GSA), the research achieves real-time holographic projection with multi-depth visualization, allowing for dynamic 3D AR-HUD displays. To ensure accurate holographic visualization, ADAS data undergoes coordinate transformation, aligning sensor-captured information with the driver’s perspective. A custom optical system, equipped with a single spatial light modulator (SLM) and spatial multiplexing, enables precise and immediate projection of road conditions. The results indicate the feasibility of using CGH-based AR-HUDs to enhance driving safety by providing an immersive interface that aligns critical ADAS information with the driver’s natural viewpoint.