Multimodal control architecture for assistive robotics

Abstract

This document present a multimodal control architecture for assistive robotics which try to minimize the possible aleatory error during the grasping process by means of visual servoing techniques. Through the gaze tracking information provided by the Tobii Pro Glasses 2 the user is capable to interact with the system in order to select the desirable object as well as indicate the intention to grasp it. At the same time, employing the 6DoF optical tracking information provided by the OptiTrack V120:Trio, the system defines the position to reach and also supervises the movement of the robot to detect some deviation in the trajectory execution.