Papers
6
Total Citations
172
H-Index
4
About
Sangjin Kim’s research spans computer vision, robotics, and energy-efficient deep learning hardware, with a focus on real-time systems and autonomous applications. In computer vision, Kim pioneered optical flow-based object tracking using non-prior training active feature models (88 citations) and developed a multiple color-filter aperture camera for depth estimation and multifocusing (38 citations), advancing computational imaging. In robotics, Kim proposed fuzzy logic-based inverse kinematics and resolved motion rate control for redundant manipulators (24 and 14 citations), offering computationally efficient alternatives to pseudo-inverse methods. More recently, Kim has led innovations in deep reinforcement learning (DRL) acceleration, introducing OmniDRL—an energy-efficient processor with dual-mode weight compression and sparse weight transposer for on-device training (2022). This work addresses critical challenges in DRL training for mobile autonomous systems, including autonomous driving and drones, by reducing energy consumption while maintaining performance. With a career bridging theoretical foundations and practical hardware design, Kim’s contributions have influenced both real-time tracking systems and next-generation edge AI processors, demonstrating sustained impact across multiple engineering domains.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3Inverse kinematics solution based on fuzzy logic for redundant manipulators24 citations · 2002
- 4Resolved motion rate control of redundant robots using fuzzy logic14 citations · 2002
- 5
- 6