Max Robotics — Global Service Robot Directory + US Certification

About

Hagit Shatkay is a leading researcher in robotics and machine learning, whose work has fundamentally advanced how autonomous systems perceive and navigate complex environments. Her primary research areas center on probabilistic modeling for robot navigation, particularly through the development and application of Hidden Markov Models (HMMs) and Partially Observable Markov Decision Processes (POMDPs). Shatkay’s major contribution lies in bridging the gap between topological and geometrical mapping. She pioneered geometrically-constrained HMMs, enabling robots to learn robust, stochastic maps of environments like office buildings and road networks without requiring strong prior structural assumptions. Her seminal 1997 paper, “Learning topological maps with weak local odometric information,” has garnered over 160 citations, establishing a foundational framework for map learning under uncertainty. This work demonstrated that even with weak odometry, topological maps—a crucial abstraction for planning—could be efficiently learned, overcoming the slow convergence and data-hungry nature of traditional Baum-Welch algorithms. By integrating geometric constraints into HMMs, Shatkay provided a principled method for robots to navigate reliably in partially observable worlds. Her research continues to influence modern autonomous systems, from service robots to self-driving vehicles, making her a pivotal figure in the evolution of intelligent navigation.