Thomas Chau
Papers
5
Total Citations
69
H-Index
5
About
Thomas Chau’s research lies at the intersection of hardware acceleration, real-time robotics, and adaptive machine learning. He is best known for pioneering work in hardware-accelerated reinforcement learning for application-specific robotic control (32 citations), demonstrating how FPGAs can enable RL agents to meet the stringent latency demands of physical systems. Chau also made foundational contributions to adaptive particle filtering, developing heterogeneous reconfigurable systems that dynamically adjust computational complexity for real-time applications (16 citations). His adaptive Sequential Monte Carlo approach (8 citations) reduced runtime overhead by intelligently varying particle set sizes, while his mapping methodology for heterogeneous FPGA-CPU platforms (8 citations) provided a systematic framework for deploying these algorithms. Notably, Chau addressed the critical challenge of real-time proximity query for dynamic active constraints in human-robot collaboration (5 citations), accelerating computationally expensive geometric calculations to enable safe, responsive physical interaction. His work consistently bridges the gap between theoretical machine learning algorithms and practical, latency-sensitive robotic systems, establishing him as a key figure in real-time adaptive computing for robotics.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3Adaptive Sequential Monte Carlo approach for real-time applications8 citations · 2012
- 4
- 5Acceleration of real-time Proximity Query for dynamic active constraints5 citations · 2013