Introduction

Path planning, obstacle avoidance, localization, and environment mapping are essential requirements of robot navigation and exploration. For micro-scale intervention and targeted therapy, disorientation due to limited microscopic field-of-view and challenges in targeting and retargeting are major hurdles to overcome. Advances in sensing, machine vision, and embedded computation have underpinned the remarkable progress of surgical robots delivering targeted therapy while negotiating complex, delicate anatomical structures. Our Centre is focussed on the fundamental theory and practical applications of robot vision and image guidance for robots used in surgery, rehabilitation and hospital automation. Areas of research include, for example, real-time SLAM and dynamic view expansion, 3D deformation recovery, probabilistic soft-tissue tracking, manifold embedding for intraluminal navigation, as well as dynamic active constraints and augmented reality visualisation. The research goal is to maximise the information content by fusing imaging/sensing with navigation to provide context aware navigation guidance, human-robot cooperative control, and in situ learning and adaption.

Research Interests

• Multisource unsupervised domain adaptation
• Robot vision with active vision and local-global mapping
• Deep learning analytics and semantic scene/image association
• Multiscale visual servoing and retargeting
• Surgical GPS and augmented reality
• XAI incorporating deep learning for robot navigation