Max Robotics — Global Service Robot Directory + US Certification

About

Kaijie Wu is a researcher whose work sits at the intersection of computer vision, robotics, and intelligent control, with a particular focus on enabling machines to perceive and interact with their environments more effectively. A central theme in Wu’s research is purposive perception—the idea that visual systems should actively seek out the best viewpoints for a given task. This is explored in their most-cited work, “Domain Adaptation for Viewpoint Estimation with Image Generation” (6 citations), which tackles the critical bottleneck of limited annotated training data for robot grasping and manipulation. Wu further advances this concept in “Model-based active viewpoint transfer for purposive perception,” addressing real-world industrial challenges like limited field-of-view and occlusion. Complementing these perceptual contributions, Wu has also made strides in physical interaction with “Improved Adaptive Variable Impedance Control for Contact Force Tracking” (3 citations), proposing a novel controller that balances stability and flexibility—a key challenge in force-sensitive tasks. Additional work in hierarchical reinforcement learning for automatic curriculum generation and multi-modal RGBD semantic segmentation for robot navigation rounds out a portfolio that consistently aims to bridge the gap between perception and action in autonomous systems.