Papers
6
Total Citations
1,427
H-Index
5
About
Kevin J. Murphy is a leading researcher in probabilistic inference, robotics, and artificial intelligence, best known for pioneering Rao-Blackwellised particle filtering (RBPF), a method that revolutionized inference in dynamic Bayesian networks (DBNs). His seminal 2001 paper on RBPF has garnered over 1,185 citations, establishing a foundational technique for handling complex, nonlinear, and non-stationary systems in fields ranging from robotics to signal processing. Murphy’s work extends to hierarchical partially observable Markov decision processes (H-POMDPs) for multi-scale robot localization, enabling efficient navigation in uncertain environments. More recently, he has focused on dynamic teleoperation, comparing joint-space and task-space mappings to allow humanoid robots to replicate human motion with agility and precision. Murphy also contributed to robotics education with HOPPY, an open-source, low-cost kit for studying dynamic legged locomotion, bridging the gap between theory and hands-on learning. His research consistently emphasizes practical, scalable solutions for real-world challenges, making him a key figure in advancing both theoretical inference methods and applied robotic systems.
Research Focus
Key Achievements
Top Papers
- 1Rao-Blackwellised Particle Filtering for Dynamic Bayesian Networks1,185 citations · 2001
- 2Rao-Blackwellised Particle Filtering for Dynamic Bayesian Networks145 citations · 2013
- 3Representing hierarchical POMDPs as DBNs for multi-scale robot localization65 citations · 2004
- 4
- 5HOPPY: An Open-source Kit for Education with Dynamic Legged Robots14 citations · 2021
- 6