Postdoc Machine Learning for Autonomous Drone Navigation

Autonomous on-board drone navigation (i.e., without human intervention) in inaccessible environments is a fundamental challenge. The goal in this project is to develop novel machine learning algorithms for autonomous drone navigation in outdoor environments including localization and synchronization for BVLOS (beyond visual line of sight) scenarios and/or GPS-denied environments, by utilizing RF signals from fixed ground stations and/or in collaboration with other drones. This includes robust holistic 3D perception through sensor fusion of on-board sensors, for e.g., lidar, radar, camera, and inertial measurement units (IMUs), algorithms for detect and avoid using advanced machine learning for object detection, object classification, and ego-motion estimation. The proposed resource-constrained algorithms will be energy-efficient and robust for in-drone perception, cognition, and control.