Project

Title: TeleGrasp – Semi-Autonomous Object Grasping using Robot Vision and Human IMU-based Teleoperation

Project Description: The 2020 COVID19 pandemic brought into the sight the need to teleoperate several objects using a robotic system from distance (e.g. safely from home) rather than having humans do that. For instance, operating in a hospital to move items around and help patients could be done through tele-presence. Workers in several (domestic, health, and industrial) environments often require to grasp and lift several objects in a tedious and repeating manner. This physical effort increases the chances of injuries either these are short or long term. Robotic platforms with arms and hands/grippers, i.e. manipulators, on mobile platforms could potentially aid humans in such scenarios towards automation of such setups (Domestic, Health, or Industry 4.0). Due to lack of purely autonomous systems that can operate in such environments, several robotic teloperative devices were introduced. These usually require special human training and tedious calibration processes, resulting in slow operations of object grasping or failures, such as object drops. This is mainly due to human-teleoperators trying to map their movement with a drifted robotic arm to position the hand/gripper exactly at the object grasping position. In this project, we propose to fuse human teleoperation, based on an IMU motion-capture system, with robot autonomy, based on visual feedback data, into a semi-autonomous system. In this way a faster teleoperation will be achieved, with more stable grasps. This is an exciting brand new area of research in robotics that is more realistic to work in real-world environments. The workers will only need to take high-level decisions, i.e. which object to grasp and where to move it, and the robot will autonomously drive the local actions related to grasping. The importance of this work lies in the fact that it can be applied to several domains that involve grasping, such as: in-house/hospital tasks, disaster scenarios or planetary exploration. We plan on controlling the robots from a distance, i.e. from London to Osaka and backwards.