HyungGi Jo
Papers
7
Total Citations
54
H-Index
5
About
HyungGi Jo is a robotics researcher specializing in autonomous navigation, simultaneous localization and mapping (SLAM), and sensor fusion for mobile robots and autonomous vehicles. His work addresses critical challenges in global localization, particularly in GPS-denied indoor environments, where traditional methods require extensive observation to converge. Jo’s most cited paper, "New Monte Carlo Localization Using Deep Initialization" (2020, 21 citations), introduces a deep learning-based initialization approach that fuses 3D LiDAR and camera data to achieve fast, accurate global localization—a significant leap over conventional Monte Carlo methods. He further advanced learning-based localization with "Mixture Density-PoseNet" (2020, 9 citations), proposing a novel neural network for monocular camera-based global localization that outperforms existing approaches. Jo has also contributed to uncertainty-aware depth networks for visual-inertial odometry (2024) and multimodal mapping using object tracking and Gaussian process regression (2024), addressing real-world dynamic environments. His earlier work on adaptive grid mapping (2013) and robust pose estimation with less reliable depth data (2015) demonstrates a sustained focus on practical, real-time robotics solutions. With a growing citation record, Jo’s research is shaping the future of autonomous navigation in complex, GPS-denied settings.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3Grid mapping adaptive to various map sizes for Sbot6 citations · 2013
- 4
- 5
- 6
- 7Visual Loop Closure Detection over Illumination Change3 citations · 2019