Homepage
Picture of Brian Wang

Hi! My name is Brian, and I'm a robotics engineer at Fleet Robotics, where I'm developing localization and mapping software that will enable an autonomous climbing robot to traverse and clean ship hulls. Previously, I worked on perception and planning for autonomous aerial vehicles at Draper.

I completed my Ph.D. degree in 2022 in the Autonomous Systems Lab at Cornell University, where I was advised by Professor Mark Campbell. My research at Cornell focused on robotic perception and planning in uncertain and unstructured environments.

Contact: bhw45 [at] cornell.edu

News

  • In January 2024, I joined Fleet Robotics as a Senior Sensor Fusion, Localization, and Mapping Engineer!

Research

[Full publications list] [Google scholar profile] [YouTube channel]

Pseudo-lidar 3-D Obstacle Detection

A pipeline for stereo camera-based 3-D object detection, using SLAM point cloud fusion to generate detector training data.

Brian H. Wang, Carlos Diaz-Ruiz, Jacopo Banfi, and Mark Campbell. "Detecting and Mapping Trees in Unstructured Environments with a Stereo Camera and Pseudo-Lidar." IEEE International Conference on Robotics and Automation (ICRA), 2021.

[pdf] [code] [video presentation]

LDLS: Label Diffusion Lidar Segmentation

Lidar point cloud instance segmentation, using camera information to avoid any need for labeled 3-D training data.

Brian H. Wang, Wei-Lun Chao, Yan Wang, Bharath Hariharan, Kilian Q Weinberger, and Mark Campbell. "LDLS: 3-D Object Segmentation Through Label Diffusion From 2-D Images." IEEE Robotics and Automation Letters (RA-L), 2019.

[pdf] [code]