Papers
17
Total Citations
648
H-Index
11
About
Kuangen Zhang is a dynamic robotics and artificial intelligence researcher whose work spans robotic assembly automation, 3D point cloud learning, and human-robot interaction for wearable systems. His most influential contributions lie in applying deep reinforcement learning to complex robotic manipulation tasks — particularly multiple peg-in-hole assembly — where he pioneered feedback deep deterministic policy gradient methods enhanced with fuzzy logic, earning over 179 and 68 citations respectively. These innovations helped overcome longstanding limitations of traditional contact-model-based control strategies by enabling robots to learn assembly skills autonomously. Zhang has also made significant strides in 3D geometric deep learning, developing the Linked Dynamic Graph CNN framework for point cloud understanding, which has accumulated nearly 200 combined citations across two publications. This work directly supports robust environmental perception for both industrial robots and wearable assistive devices. His research on human locomotion intent prediction — through unsupervised cross-subject adaptation and gaze-fused foot placement prediction — demonstrates a compelling commitment to advancing exoskeleton and assistive robot control. His COVID-19-inspired wheel-legged robotic limb further highlights his ability to translate fundamental research into timely real-world applications, cementing his reputation as a versatile and impactful contributor to intelligent robotics.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5Unsupervised Cross-Subject Adaptation for Predicting Human Locomotion Intent49 citations · 2020
- 6
- 7
- 8
- 9
- 10Directional PointNet: 3D Environmental Classification for Wearable Robotics14 citations · 2019