Hidehiko Hosaka

Papers

Total Citations

H-Index

About

Hidehiko Hosaka is a leading researcher in field robotics, specializing in autonomous navigation and perception for unstructured, rough-terrain environments. His core research areas encompass LiDAR-based localization, 3D scan matching, and deep learning for terrain detection, with a primary focus on enabling mobile robots and heavy forestry vehicles to operate safely in forests and on unpaved roads. Hosaka’s major contributions include developing a forest road surface detection system that combines LiDAR-SLAM with a U-Net convolutional neural network, allowing robots to dynamically identify passable surfaces despite changing geometry, lighting, and surface conditions. He has also advanced 3D scan matching techniques for mobile robot localization over rough terrain, addressing the significant odometry errors that occur when traveling on uneven ground. His work on robust, high-speed localization methods has been cited over 10 times in the field, demonstrating its value for autonomous forestry and off-road applications. Through these innovations, Hosaka is helping to automate labor-intensive forestry tasks, improving safety and efficiency in one of the most challenging environments for autonomous systems.