Max Robotics — Global Service Robot Directory + US Certification

About

Mark Cummins is a robotics and computer vision researcher whose work has made landmark contributions to the field of autonomous navigation and place recognition. He is best known for developing FAB-MAP (Fast Appearance-Based Mapping), a probabilistic framework for Simultaneous Localization and Mapping (SLAM) that enables mobile robots to recognize previously visited locations purely from visual appearance data — without relying on GPS or infrastructure. Introduced in 2007–2008, FAB-MAP has become a cornerstone reference in appearance-based navigation, amassing over 1,475 citations and establishing Cummins as a leading figure in scalable, real-world robot localization. His research elegantly addresses one of robotics' most challenging problems: how a robot can distinguish between truly new environments and revisited ones, even when many locations look deceptively similar. By learning generative models of visual appearance and applying rigorous probabilistic inference, Cummins brought both mathematical elegance and practical efficiency to large-scale mapping. His work on accelerated SLAM using probabilistic bail-out conditions further demonstrated his focus on computational tractability. Beyond navigation, Cummins also contributed to semantic urban mapping, developing probabilistic frameworks for labeling city environments using spatial and temporal context — work that bridges autonomous robotics and intelligent scene understanding.