Papers
13
Total Citations
217
H-Index
7
About
Zexi Chen is a leading researcher in robotics, with a primary focus on LiDAR-based global localization, place recognition, and pose estimation for autonomous navigation. Their most impactful contribution is the development of **DiSCO** (Differentiable Scan Context with Orientation), a seminal work with 103 citations that introduced a differentiable approach to robustly retrieve a query scan from a map database, significantly advancing the field of appearance-invariant place recognition. Building on this, Chen’s **One RING to Rule Them All** (24 citations) further unified place recognition with orientation and translation estimation using Radon sinograms, while **DPCN** and **DPCN++** pioneered end-to-end differentiable phase correlation networks for heterogeneous sensor matching and versatile pose registration. Beyond localization, Chen has made notable contributions to robot learning, including kinematic motion retargeting for sign language (26 citations) and hierarchical imitation learning for autonomous driving (21 citations), as well as dynamic manipulation for in-flight object catching. Their work has been published in top venues like *IEEE Robotics and Automation Letters* and *ICRA*, and they were a key member of the champion team ZJUNLict at RoboCup 2019. With a total citation count exceeding 200, Chen’s research is defining new standards for robust, learning-based robot perception and control.
Research Focus
Key Achievements
Top Papers
- 1DiSCO: Differentiable Scan Context With Orientation103 citations · 2021
- 2
- 3
- 4
- 5Neural Motion Prediction for In-flight Uneven Object Catching12 citations · 2021
- 6
- 7
- 8
- 9ZJUNlict Extended Team Description Paper for RoboCup 20194 citations · 2019
- 10