Paper accepted by IEEE Transactions on Radar Systems

Our paper “Improving Height Estimation for Stationary Targets with 3D Automotive Radar: From Uncertainty Analysis to Temporal Filtering” has been accepted to IEEE Transactions on Radar Systems! 🎉

In this work, we tackle one of the key challenges in 3D automotive radar: accurate height estimation for stationary targets. By combining rigorous uncertainty analysis with a robust temporal filtering approach (EKF + data association), we significantly improve radar-based 3D scene understanding without requiring additional hardware.

✅ Highway experiments: height accuracy within 1m improved from 53.41% → 62.32%
✅ Campus experiments: height accuracy within 1m improved from 47.74% → 57.56%
✅ A cost-effective way to bridge the gap between standard 3D radars and expensive 4D radars.

If you’re interested in radar perception, autonomous driving, or uncertainty-aware filtering, check out our preprint here: 🔗 Preprint on TechRxiv

Big thanks to my co-authors Prof. Chieh-Chih Wang and Prof. Wen-Chieh Lin for their guidance and collaboration. 🙌