arXiv subjects

Tsetserukou Dzmitry

Publications and source records attributed to Tsetserukou Dzmitry.

3 recordsLinked to original sources

arXiv2022

CobotTouch: AR-based Interface with Fingertip-worn Tactile Display for Immersive Operation/Control of Collaborative Robots

Complex robotic tasks require human collaboration to benefit from their high dexterity. Frequent human-robot interaction is mentally demanding and time-consuming. Intuitive and easy-to-use robot control interfaces reduce the negative influence on workers, especially inexperienced users. In this paper, we present CobotTouch, a novel intuitive robot control interface with fingertip haptic feedback. The proposed interface consists of a projected Graphical User Interface on the robotic arm to control the position of the robot end-effector based on gesture recognition, and a wearable haptic interface to deliver tactile feedback on the user's fingertips. We evaluated the user's perception of the designed tactile patterns presented by the haptic interface and the intuitiveness of the proposed system for robot control in a use case. The results revealed a high average recognition rate of 75.25\% for the tactile patterns. An average NASA Task Load Index (TLX) indicated small mental and temporal demands proving a high level of the intuitiveness of CobotTouch for interaction with collaborative robots.

cs.HC↗

arXiv2022

DeepXPalm: Tilt and Position Rendering using Palm-worn Haptic Display and CNN-based Tactile Pattern Recognition

Telemanipulation of deformable objects requires high precision and dexterity from the users, which can be increased by kinesthetic and tactile feedback. However, the object shape can change dynamically, causing ambiguous perception of its alignment and hence errors in the robot positioning. Therefore, the tilt angle and position classification problem has to be solved to present a clear tactile pattern to the user. This work presents a telemanipulation system for plastic pipettes consisting of a multi-contact haptic device LinkGlide to deliver haptic feedback at the users' palm and two tactile sensors array embedded in the 2-finger Robotiq gripper. We propose a novel approach based on Convolutional Neural Networks (CNN) to detect the tilt and position while grasping deformable objects. The CNN generates a mask based on recognized tilt and position data to render further multi-contact tactile stimuli provided to the user during the telemanipulation. The study has shown that using the CNN algorithm and the preset mask, tilt, and position recognition by users is increased from 9.67% using the direct data to 82.5%.

cs.RO↗

arXiv2021

CobotAR: Interaction with Robots using Omnidirectionally Projected Image and DNN-based Gesture Recognition

Several technological solutions supported the creation of interfaces for Augmented Reality (AR) multi-user collaboration in the last years. However, these technologies require the use of wearable devices. We present CobotAR - a new AR technology to achieve the Human-Robot Interaction (HRI) by gesture recognition based on Deep Neural Network (DNN) - without an extra wearable device for the user. The system allows users to have a more intuitive experience with robotic applications using just their hands. The CobotAR system assumes the AR spatial display created by a mobile projector mounted on a 6 DoF robot. The proposed technology suggests a novel way of interaction with machines to achieve safe, intuitive, and immersive control mediated by a robotic projection system and DNN-based algorithm. We conducted the experiment with several parameters assessment during this research, which allows the users to define the positives and negatives of the new approach. The mental demand of CobotAR system is twice less than Wireless Gamepad and by 16\% less than Teach Pendant.

cs.RO↗