Tsung-Wun (David) Wang

Tsung-Wun (David) Wang

Student at National Taiwan University

Research Contributions

My master’s thesis can be categorized into three parts, from 3D reconstruction to 2D and 3D navigation of AMRs.

  • Title: Improved Real-Time Dense ORB SLAM with GPU Implementation
    • Authors: Tsung-Wun Wang, Han-Pang Huang, Chiou-Shann Fuh
    • Status: Accepted (August 2023) by the 36th IPPR Conference on Computer Vision, Graphics, and Image Processing, Kinmen, Taiwan.
    • Contribution: This paper presents a GPU-accelerated RGB-D SLAM module that enhances real-time performance and depth measurement quality, crucial for effective 3D reconstruction in robotics.
  • Title: Search-based Path Replanning for Autonomous Navigation and Obstacle Avoidance
    • Authors: Tsung-Wun Wang, Han-Pang Huang
    • Status: Rejected (October 2023) by Advanced Robotics (RSJ).
    • Contribution: This paper analyzes search-based pathfinding algorithms for 2D navigation, enhancing real-time obstacle avoidance strategies in dynamic settings.
  • Title: Vision-based Autonomous Robot Navigation in 3D Dynamic Environments
    • Authors: Tsung-Wun Wang, Han-Pang Huang
    • Status: Rejected (November 2023) by IEEE Transactions on Automation Science and Engineering.
    • Contribution: This work introduces a comprehensive navigation system for mobile robots that integrates SLAM with dynamic obstacle avoidance, addressing real-time challenges in 3D environments.


In addition to the above-mentioned research fields, GMapping and navigation mesh (NavMesh) are also of interest to me. For example, I

  • Conducted two-dimensional (2D) LiDAR SLAM and presented modified GMapping that addresses loop-closing.
  • Developed different pathfinding algorithms on navigation mesh (NavMesh) and implemented the path replanning mechanism.

Although two of the journal papers were recently rejected, I view this as an opportunity for further refinement and development. If you are interested in discussing any of these topics or my research in more detail, please feel free to contact me directly via email. Thank you!