Tianshu Wu
Papers
2
Total Citations
12
H-Index
2
About
Tianshu Wu is a robotics researcher pushing the boundaries of generalizable manipulation and intelligent disassembly. His work centers on bridging the gap between high-level reasoning and precise physical interaction, with a particular focus on object-centric manipulation and compliant control strategies. Wu’s most impactful contribution, “OmniManip” (2025, 9 citations), tackles the fundamental challenge of enabling robots to operate in unstructured environments by introducing object-centric interaction primitives as spatial constraints. This approach leverages Vision-Language Models for commonsense reasoning while compensating for their lack of fine-grained 3D spatial understanding—a critical step toward truly general robotic manipulation. In complementary work, Wu developed a “Peg-Hole Robotic Disassembly Compliant Strategy based on Soft Actor-Critic Algorithm” (2024, 3 citations), addressing the practical industrial challenge of remanufacturing end-of-life products. By applying reinforcement learning to manage disassembly forces, he offers an effective solution for improving remanufacturing efficiency. Together, these contributions demonstrate Wu’s commitment to creating robots that can both understand and act in complex, real-world environments, marking him as a rising talent in the field of robotic manipulation and intelligent automation.
Research Focus
Key Achievements
Top Papers
- 1
- 2