5DT Data Glove Research Articles

Deep learning‐based grasp‐detection method for a five‐fingered industrial robot hand

To improve the accuracy of robotic grasp in some uncertain environments, a deep learning‐based object‐detection method for a five‐fingered industrial robot hand model is proposed in this study. The authors first design a five‐fingered industrial robot hand model with 21‐degrees of freedom (DOF). Based on the sensor data of a 5DT data glove, the industrial robot hand can be controlled in real time. They use the object‐detection network's faster regions convolutional neural network and single shot multibox detector to locate the grasp objects. To optimise the robotic grasp detection, two grasp‐predictor methods, direct grasp predictor and multi‐modal grasp predictor, are applied to obtain the best graspable region. In the simulation designed in this study, cooperating with a 6‐DOF robot arm, the five‐fingered industrial robot hand can detect an object accurately and grasp it steadily.

Development and Design Wearable Sensorial System of Low Cost to Determine the Metacarpophalangeal and Interphalangeal Joints Angles

This paper presents the results of a low-cost wearable sensory system design process for the simultaneous capture of the angles of the metacarpophalangeal and interphalangeal joints of the human hand, as well as the characterization of its performance compared with similar commercial systems, The processing subsystem has machine learning algorithms that were compared in order to identify the optimal technique for capturing information corresponding to basic movements such as hold different objects and characterizing displacements at the level of phalanges in the hand. The development was carried out at the hardware level through the manufacture of a glove with flexible resistive sensors of the reference SEN-10264 (it is a sensor with resistive cells that varies its value in ohms depending on the flexion to which it is subjected) two per each finger, for a total of 10 sensors adhered to the surface of the glove, after going through a coupling stage (to avoid large amounts of noise) the signals generated by the sensors are acquired through the data capture module Analogs of a Microcontroller. In terms of software, the Microcontroller was first programmed with a digital filter stage for the signals acquired by the SEN-10264 flex sensor sensors and a machine learning algorithm for the processing of these and in this way obtain the individual metacarpophalangeal and interphalangeal joints angles of the hand evaluated. Finally, obtain the sensor data in terms of angles and export it to a flat text file to be graphed. For a total of 10 people, multiple acquisitions were made with the developed system and in this way, similar patterns were determined in the angles of the metacarpophalangeal and interphalangeal between the test subjects in basic tasks such as hold different types of objects. Finally, the development of the sensor system was validated with two types of wearable sensors: the 5DT Data Glove device and the Optitrack Flex 3 high-speed cameras, these two types of devices have acquisition protocols predetermined by their manufacturers. Given that the variables to obtain are the metacarpophalangeal and interphalangeal joints angles, the first 5DT Data Glove device delivers only one output signal for each finger of the evaluated hand, basically the software of the manufacturer performs a mathematical interpolation and approximations depending on the postures of the fingers, which makes it incapable by means of this technique to determine if a flexion is occurring from the metacarpophalangeal and interphalangeal joints separately, the Optitrack Flex 3 high-speed cameras generate real-time data graphs of the position in 3d and the trajectories of each of the joints of the hand, the problem is that the system performs the capture of data through markers located in the hand and first of all the size of these makes them crash in a task of grip in addition to making the task of characterizing the ag arrest because the viewing angles of the cameras must detect each marker located in the joints of the hand. The individual angles will be used later to carry out a simulation project of a postural biological synergy of some basic object grabs (with a control command generating movement in several degrees of freedom) through the implementation of an artificial intelligence algorithm.

An evaluation of asymmetric interfaces for bimanual virtual assembly with haptics

Immersive computing technology provides a human---computer interface to support natural human interaction with digital data and models. One application for this technology is product assembly methods planning and validation. This paper presents the results of a user study which explores the effectiveness of various bimanual interaction device configurations for virtual assembly tasks. Participants completed two assembly tasks with two device configurations in five randomized bimanual treatment conditions (within subjects). A Phantom Omni® with and without haptics enabled and a 5DT Data Glove were used. Participant performance, as measured by time to assemble, was the evaluation metric. The results revealed that there was no significant difference in performance between the five treatment conditions. However, half of the participants chose the 5DT Data Glove and the haptic-enabled Phantom Omni® as their preferred device configuration. In addition, qualitative comments support both the preference of haptics during the assembly process and comments confirming Guiard's kinematic chain model.

Robust Hand Motion Tracking through Data Fusion of 5DT Data Glove and Nimble VR Kinect Camera Measurements.

Vision based interfaces for human computer interaction have gained increasing attention over the past decade. This study presents a data fusion approach of the Nimble VR vision based system, using the Kinect camera, with the contact based 5DT Data Glove. Data fusion was achieved through a Kalman filter. The Nimble VR and filter output were compared using measurements performed on (1) a wooden hand model placed in various static postures and orientations; and (2) three differently sized human hands during active finger flexions. Precision and accuracy of joint angle estimates as a function of hand posture and orientation were determined. Moreover, in light of possible self-occlusions of the fingers in the Kinect camera images, data completeness was assessed. Results showed that the integration of the Data Glove through the Kalman filter provided for the proximal interphalangeal (PIP) joints of the fingers a substantial improvement of 79% in precision, from 2.2 deg to 0.9 deg. Moreover, a moderate improvement of 31% in accuracy (being the mean angular deviation from the true joint angle) was established, from 24 deg to 17 deg. The metacarpophalangeal (MCP) joint was relatively unaffected by the Kalman filter. Moreover, the Data Glove increased data completeness, thus providing a substantial advantage over the sole use of the Nimble VR system.

Powered exoskeleton with palm degrees of freedom for hand rehabilitation.

Robotic rehabilitation is a currently underutilised field with the potential to allow huge cost savings within healthcare. Existing rehabilitation exoskeletons oversimplify the importance of movement of the hand while undertaking everyday tasks. Within this study, an investigation was undertaken to establish the extent to which the degrees of freedom within the palm affect ability to undertake everyday tasks. Using a 5DT data glove, bend sensing resistors and restrictors of palm movement, 20 participants were recruited to complete tasks that required various hand shapes. Collected data was processed and palm arching trends were identified for each grasping task. It was found that the extent of utilizing arches in the palm varied with each exercise, but was extensively employed throughout. An exoskeleton was subsequently designed with consideration of the identified palm shapes. This design included a number of key features that accommodated for a variety of hand sizes, a novel thumb joint and a series of dorsally mounted servos. Initial exoskeleton testing was undertaken by having a participant complete the same exercises while wearing the exoskeleton. The angles formed by the user during this process were then compared to those recorded by 2 other participants who had completed the same tasks without exoskeleton. It was found that the exoskeleton was capable of forming the required arches for completing the tasks, with differences between participants attributed to individual ergonomic differences.

Integration and evaluation of haptic feedbacks: from CAD models to virtual prototyping

This paper studies the benefits that haptic feedback can provide in performing complex maintenance tasks using virtual mock-ups. We have carried out user study that consisted on two experiments where participants had to perform an accessibility task. A human-scale string-based haptic interface was used to provide the operator with haptic stimuli. A prop was used to provide grasp feedback. A mocap system tracks user’s hand and head movements while a 5DT data-glove is used to measure finger flexion. In the first experiment the effect of haptic (collision) and visual feedback are investigated. In the second experiment we investigated the effect of haptic guidance on operator performance. The results were analyzed in terms of task completion time and collision avoidance. Experiments show that haptic stimuli proved to be more efficient than visual ones. In addition, haptic guidance helped the operators to correct trajectories and hence improve their performance.

APLIKASI TIGA DIMENSI VIRTUAL DENGAN MENGGUNAKAN 5DT DATA GLOVE 5 ULTRA

One of active area in information technology research is development of virtual reality (VR). It is a technology which can simulate real world activities in a virtual world. Virtual reality consists of software and hardware which is usually in the form of device that captures human movement. Then VR sends signal of that movement to computer. The signal will be received and processed and shown on the screen, so that the movement can be simulated. In this research, we develop an application that can receive hand movement input with 5DT Data Glove 5 Ultra and create natural and accurate hand movement animation output for hand opening closing and other gestures. Experiments show that there are some gestures though have different physical shapes but are recognized as the same gestures because of sensor limitation in data glove.

Collision Detection and Part Interaction Modeling to Facilitate Immersive Virtual Assembly Methods

Realistic part interaction is an important component of an effective virtual assembly application. Both collision detection and part interaction modeling are needed to simulate part-to-part and hand-to-part interactions. This paper examines several polygonal-based collision detection packages and compares their usage for virtual assembly applications with the Voxmap PointShell (VPS) software developed by the Boeing Company. VPS is a software developer’s toolkit for real-time collision and proximity detection, swept-volume generation, dynamic animation, and 6 degree-of-freedom haptics which is based on volumetric collision detection and physically based modeling. VPS works by detecting interactions between two parts: a dynamic object moving in the virtual environment, and a static object defined as a collection of all other objects in the environment. The method was found to provide realistic collision detection and physically-based modeling interaction, with good performance at the expense of contact accuracy. Results from several performance tests on VPS are presented. This paper concludes by presenting how VPS has been implemented to handle multiple dynamic part collisions and two-handed assembly using the 5DT dataglove in a projection screen virtual environment.