Papers
5
Total Citations
64
H-Index
4
About
Siwei Lyu is a leading researcher in robotic manipulation and state estimation, with a focus on enabling robots to physically interact with their environments through contact-aware perception and control. His work bridges the gap between dynamic systems theory and practical robotic grasping, particularly through the development of C-SLAM (Contact Simultaneous Localization and Modeling), a framework that allows robots to simultaneously track object motion and contact locations during manipulation. Lyu has made foundational contributions to contact-aware state estimation (CASE), using particle filters and complementarity-based dynamics to accurately estimate object pose and velocity during active manipulation—even under intermittent contact. His research has been cited over 60 times, with his 2013 paper on dynamic Bayesian approaches to grasp acquisition receiving 29 citations. More recently, Lyu has ventured into deep reinforcement learning for robotic control, introducing RMBench, a benchmarking platform for evaluating RL algorithms in manipulator tasks. He has also explored industrial applications, such as laser vision-based welding seam tracking for curved surfaces. Lyu’s work is essential reading for anyone interested in robust, physically grounded robot manipulation.
Research Focus
Key Achievements
Top Papers
- 1
- 2State estimation for dynamic systems with intermittent contact17 citations · 2015
- 3A comparative study of contact models for contact-aware state estimation10 citations · 2015
- 4
- 5