2018/09/21 by Ali Shafti, Shafti, Ali, Pavel Orlov +3 · 3 citations
Computer Science · Engineering · #Gaze Tracking and Assistive Technology #Robot Manipulation and Learning #Soft Robotics and Applications
paper · pdf · doi:10.48550/arxiv.1809.08095
Assistive robotic systems endeavour to support those with movement\ndisabilities, enabling them to move again and regain functionality. Main issue\nwith these systems is the complexity of their low-level control, and how to\ntranslate this to simpler, higher level commands that are easy and intuitive\nfor a human user to interact with. We have created a multi-modal system,\nconsisting of different sensing, decision making and actuating modalities,\nleading to intuitive, human-in-the-loop assistive robotics. The system takes\nits cue from the user's gaze, to decode their intentions and implement\nlow-level motion actions to achieve high-level tasks. This results in the user\nsimply having to look at the objects of interest, for the robotic system to\nassist them in reaching for those objects, grasping them, and using them to\ninteract with other objects. We present our method for 3D gaze estimation, and\ngrammars-based implementation of sequences of action with the robotic system.\nThe 3D gaze estimation is evaluated with 8 subjects, showing an overall\naccuracy of 4.68\±0.14cm. The full system is tested with 5 subjects, showing\nsuccessful implementation of 100 % of reach to gaze point actions and full\nimplementation of pick and place tasks in 96 %, and pick and pour tasks in\n76 % of cases. Finally we present a discussion on our results and what future\nwork is needed to improve the system.\n