Jun Zhang

Papers

Total Citations

208

H-Index

About

Jun Zhang is a researcher whose work spans autonomous driving, robotics perception, and intelligent vehicle systems. His most notable contribution is the development of **OverlapTransformer**, a transformer-based neural network architecture designed for LiDAR-based place recognition in autonomous vehicles. Published in 2022, this work addresses one of the core challenges in autonomous navigation — enabling vehicles to recognize previously visited locations accurately and efficiently, even under varying yaw angles. The method is particularly valuable for loop closure detection in Simultaneous Localization and Mapping (SLAM) and global localization tasks, and has garnered an impressive 196 citations, reflecting its significant uptake within the robotics and autonomous driving communities. Beyond deep learning for perception, Zhang has also contributed to the field of intelligent vehicle control, exploring the automation of shift-scheduling calibration through bionic optimization techniques such as particle swarm optimization combined with personalized driver modeling. This earlier work demonstrates his broader interest in making vehicle systems smarter and more adaptive without heavy human intervention. Together, Zhang's research highlights a consistent focus on reducing human effort in complex vehicular tasks — whether through learned perception systems or automated calibration pipelines — making him a versatile contributor to the advancement of intelligent and autonomous transportation technologies.