Max Robotics — Global Service Robot Directory + US Certification

About

Marco Scoffier is a researcher whose work sits at the intersection of autonomous robotics, computer vision, and machine learning, with a particular focus on enabling mobile robots to navigate complex off-road environments. His most influential contribution, "Learning Long-Range Vision for Autonomous Off-Road Driving" (2009), has garnered over 316 citations and introduced a self-supervised learning framework capable of classifying terrain at distances extending to the horizon — a significant leap beyond the short-range limitations of conventional stereo vision systems. Building on this, his 2008 work applying deep belief networks to long-range vision (87 citations) was an early and prescient use of deep learning for robotic perception, predating much of the field's mainstream adoption of such techniques. Scoffier further advanced the field through research on multi-range navigation architectures, probabilistic mapping under uncertainty, and data-driven vehicle dynamics modeling via maneuver dictionaries. Much of this work was developed and validated on the DARPA LAGR mobile robot platform, giving his contributions strong real-world grounding. Collectively, his research helped establish self-supervised, learning-based perception as a viable and powerful paradigm for autonomous outdoor navigation, influencing subsequent generations of autonomous vehicle and field robotics research.