Builder · Learner · Researcher · Engineer · Coder
Research Associate at Nanyang Technological University, working at the Centre for Advanced Robotics Technology Innovation (CARTIN) on resilient SLAM, multi-sensor fusion, thermal-radar perception, and 3D reconstruction for autonomous navigation in perception-degraded environments.
Scroll for ProjectsResearch and engineering projects where I made primary technical contributions across dense mapping, robust SLAM, sensor synchronization, and degradation-aware fusion.
Thermal Dense Mapping
Fireground-oriented dense mapping from thermal imagery, built on a thermal-radar-inertial SLAM backbone for smoke, water spray, fire, darkness, and other perception-degraded environments. The current prototype uses odometry-guided Depth Anything 3 inference and reaches about 0.5 Hz online dense reconstruction on an RTX 5090 laptop.
Robust SLAM
Real-time stereo thermal-radar-inertial localization for fireground-style environments, validated across smoke, water mist, fire, warehouse, indoor-outdoor, and semi-outdoor trials, with normalized APE below 1% of trajectory length on ground-truth sequences.
Sensor Sync
A hardware-software synchronization stack that aligns stereo thermal cameras, IMU, radar, and robot data through PTP network time, MCU-generated trigger pulses, and driver-level timestamp assignment for repeatable multi-sensor data collection.
Degradation-Aware SLAM
A RA-L paper and open-source SLAM framework that adaptively fuses LiDAR, thermal imagery, and inertial measurements for robust state estimation in smoke, darkness, and other perception-degraded environments.
Additional collaborative projects with focused contributions across fireground perception, robotic teleoperation, and published thermal odometry.
Fireground BEV Occupancy
Thermal-radar occupancy prediction for degraded visibility, achieving about 40 mIoU in perception-degraded scenes while running as a 1 Hz prototype on an RTX 2080 Ti.
Thermal Odometry
A RA-L paper on real-time monocular thermal-inertial odometry that uses thermal edge structures, adaptive DT-KLT tracking, and sliding-window optimization for localization under darkness, smoke, and weak visible texture.
VR Gimbal
A teleoperated pitch-yaw sensor gimbal for mobile robots, combining mechanical design, embedded real-time control, VR interaction, and onboard multi-sensor SLAM support for intuitive viewpoint control.
Peer-reviewed work on thermal, LiDAR, and inertial odometry for robust robotic perception.
IEEE Robotics and Automation Letters · 2025
Y. Wang*, Y. Liu*, L. Chen, H. Chen, and S. Zhang
IEEE Robotics and Automation Letters · 2023
Y. Wang, H. Chen, Y. Liu, and S. Zhang
I am a Research Associate at NTU CARTIN, supervised by Prof. Danwei Wang. My work focuses on building deployable perception and SLAM systems for robots operating in smoke, darkness, weak texture, and other degraded sensing conditions.
I received my M.Sc. in Computer Control and Automation from Nanyang Technological University, supervised by Prof. Danwei Wang and Dr. Jun Zhang, and my B.Eng. in Automation from Harbin Institute of Technology (Shenzhen), supervised by Prof. Haoyao Chen and Dr. Yu Wang.
ROS · SLAM · Sensor Fusion · Field Deployment
Thermal Camera · LiDAR · 4D Radar · IMU
Dense Mapping · BEV Occupancy · 3D Reconstruction
C++ · Python · Embedded Integration · Timestamp Sync