Max Robotics — Global Service Robot Directory + US Certification

About

Alper Nabi Erkan’s research lies at the intersection of autonomous navigation, machine learning, and computer vision, with a particular focus on enabling robots to perceive and traverse complex, unstructured off-road environments. His most influential work, "Deep belief net learning in a long-range vision system for autonomous off-road driving" (87 citations), pioneered the use of deep belief networks for long-range terrain classification, allowing autonomous vehicles to perform high-level strategic planning by accurately interpreting the landscape up to the horizon. This was complemented by his work on online learning for off-road robots (30 citations), which introduced a spatial label propagation method to train classifiers in real-time, enabling robots to predict traversability across entire scenes using sparse stereo data. Erkan also made notable contributions to semi-supervised learning, developing a generalized maximum entropy framework (23 citations) that expanded the theoretical foundations of the field. His research on learning probabilistic models of grasp affordances under limited supervision (15 citations) further demonstrated his versatility, addressing the challenge of robotic manipulation with minimal labeled data. Erkan’s work has been instrumental in advancing autonomous off-road navigation, earning him recognition for combining theoretical rigor with practical, real-world applications.