• International conference on construction engineering and project management
• /
• 2022.06a
• /
• pp.328-335
• /
• 2022

With the advance of robot capabilities and functionalities, construction robots assisting construction workers have been increasingly deployed on construction sites to improve safety, efficiency and productivity. For close-proximity human-robot collaboration in construction sites, robots need to be aware of the context, especially construction worker's behavior, in real-time to avoid collision with workers. To recognize human behavior, most previous studies obtained 3D human poses using a single camera or an RGB-depth (RGB-D) camera. However, single-camera detection has limitations such as occlusions, detection failure, and sensor malfunction, and an RGB-D camera may suffer from interference from lighting conditions and surface material. To address these issues, this study proposes a novel method of 3D human pose estimation by extracting 2D location of each joint from multiple images captured at the same time from different viewpoints, fusing each joint's 2D locations, and estimating the 3D joint location. For higher accuracy, the probabilistic representation is used to extract the 2D location of the joints, considering each joint location extracted from images as a noisy partial observation. Then, this study estimates the 3D human pose by fusing the probabilistic 2D joint locations to maximize the likelihood. The proposed method was evaluated in both simulation and laboratory settings, and the results demonstrated the accuracy of estimation and the feasibility in practice. This study contributes to ensuring human safety in close-proximity human-robot collaboration by providing a novel method of 3D human pose estimation.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2007.11a
• /
• pp.159-162
• /
• 2007

The inverse kinematics problem in robotics is an essential work for grasping and manipulation tasks by robotic and humanoid hands. In this paper, an intelligent neural learning scheme for solving such inverse kinematics of humanoid fingers is presented. Specifically, a multi-layered neural network is utilized for effective inverse kinematics, where a dynamic neural learning algorithm is employed. Also, a bio-mimetic feature of general human fingers is incorporated to the learning scheme. The usefulness of the proposed approach is verified by simulations.

• 제어로봇시스템학회:학술대회논문집
• /
• 2001.10a
• /
• pp.44.3-44
• /
• 2001

This articles describes a novel design elastic joint manipulator for a mobile robot, which works in an office environment with humans. The primary goal of this manipulator design is safeness on collision and contact. To achieve this, each joint is made of an elastic element and this is driver with a high ratio gear tram. The performance was verified, however, it has a serious drawback. It produce vibration, due to the elastic joints and high ratio gear train. We found that a sliding mode controller has an excellent performance for reducing such vibration. Results of computer simulation and experiments are shown.

• 제어로봇시스템학회:학술대회논문집
• /
• 2002.10a
• /
• pp.78.6-78
• /
• 2002

$\textbullet$ The Gifu Hand III is a 5-fingered hand driven by built-in servomotors and has 20 joints with 10 DOF. $\textbullet$ The backlash of transmission, the mobility space, and the opposability of the thumb are improved. $\textbullet$ The new distributed tactile sensor with 859 detecting points is mounted on the hand surface. $\textbullet$ Experiments of grasping objects by a grasping strategy imitating human grasping reflex are shown.

• Journal of Korea Multimedia Society
• /
• v.10 no.6
• /
• pp.716-725
• /
• 2007

In this paper we describe a motion tracking algorithm for 3D human animation using stereo vision system. This allows us to extract the motion data of the end effectors of human body by following the movement through segmentation process in HIS or RGB color model, and then blob analysis is used to detect robust shape. When two hands or two foots are crossed at any position and become disjointed, an adaptive algorithm is presented to recognize whether it is left or right one. And the real motion is the 3-D coordinate motion. A mono image data is a data of 2D coordinate. This data doesn't acquire distance from a camera. By stereo vision like human vision, we can acquire a data of 3D motion such as left, right motion from bottom and distance of objects from camera. This requests a depth value including x axis and y axis coordinate in mono image for transforming 3D coordinate. This depth value(z axis) is calculated by disparity of stereo vision by using only end-effectors of images. The position of the inner joints is calculated and 3D character can be visualized using inverse kinematics.

• Proceedings of the Korean Society of Marine Engineers Conference
• /
• 2011.10a
• /
• pp.97-98
• /
• 2011

This paper is about the study of a new exo-skeleton device applying a gravity compensator. The exo-skeleton devices is to reduce the external torque applied to the human body joint for the purpose of helping the disabled, reducing heavy payload for industry workers or military soldiers. Most of the exoskeleton devices are actuated by the motors, but motors are limited in energy such that a short durability is always a big problem. In this paper, an exoskeleton device using a new gravity compensator based on a torsion bar is proposed to reduce the torque load applied to human body joints. The exoskeleton device is designed for the lower body of human. Analyses on the torsion bar spring and link of the exoskeleton device using FEM method were performed. To reduce the applied torque to the human joint, a torsion bar gravity compensator is applied to the exoskeleton. The effect of the torsion bar compensator for the exoskeleton device was verified through load test using developed test equipment.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2002.05a
• /
• pp.752-759
• /
• 2002

This paper presented a design and a control of a biped walking RGO and walking simulation by this system. The biped walking RGO was distinguished from the other one by which had a very light-weight and a new RGO type with 12-servo motors. The vibration evaluation of the dynamic PLS on the biped walking RGO was used to access by the 3-axis accelerometer with a low frequency vibration for the spinal cord injuries. The gait of a biped walking RGO depended on the constrains of mechanical kinematics and the initial posture. The stability of dynamic walking was investigated by a ZMP (Zero Moment Point) of the biped walking RGO. It was designed according to a human wear type and was able to accomodate itself to a human environments. The joints of each leg were adopted with a good kinematic characteristics. To test of the analysis of joint kinematic properties, we did the strain stress analysis of the dynamic PLS and the analysis study of FEM with a dynamic PLS. It will be expect that the spinal cord injury patients are able to recover effectively with a biped walking RGO.

• The Journal of Korea Robotics Society
• /
• v.17 no.3
• /
• pp.322-333
• /
• 2022

The Avatar robot, which is one of the teleoperation robots, aims to enable users to feel the robot as a part of the body to intuitively and naturally perform various tasks. Considering the purpose of the avatar robot, an end-effector identical to a human hand is advantageous, but a robotic hand with human hand level performance has not yet been developed. In this paper we propose a new 3-finger robotic hand with human-avatar hand posture mapping algorithm which were integrated with TOCABI-AVATAR, one of the teleoperation system. Due to the flexible rolling contact joints and tendon driven mechanism applied to the finger, the finger could implement adaptive grasping and absorb the impact force caused by unexpected contacts. In addition, human-avatar hand mapping algorithm using five calibration hand postures propose to compensate physical differences between operators. Using the TOCABI-AVATAR system with the robotic hands and mapping algorithm, the operator can perform 13 out of 16 hand postures of grasping taxonomy and 4 gestures. In addition, using the system, we participated in the ANA AVATAR XPRIZE Semi-final and successfully performed three scenarios which including various social interactions as well as object manipulation.

• ETRI Journal
• /
• v.44 no.2
• /
• pp.286-299
• /
• 2022

Human activity recognition in real time is a challenging task. Recently, a plethora of studies has been proposed using deep learning architectures. The implementation of these architectures requires the high computing power of the machine and a massive database. However, handcrafted features-based machine learning models need less computing power and very accurate where features are effectively extracted. In this study, we propose a handcrafted model based on three-dimensional sequential skeleton data. The human body skeleton movement over a frame is computed through joint positions in a frame. The joints of these skeletal frames are projected into two-dimensional space, forming a "movement polygon." These polygons are further transformed into a one-dimensional space by computing amplitudes at different angles from the centroid of polygons. The feature vector is formed by the sampling of these amplitudes at different angles. The performance of the algorithm is evaluated using a support vector machine on four public datasets: MSR Action3D, Berkeley MHAD, TST Fall Detection, and NTU-RGB+D, and the highest accuracies achieved on these datasets are 94.13%, 93.34%, 95.7%, and 86.8%, respectively. These accuracies are compared with similar state-of-the-art and show superior performance.

• Science of Emotion and Sensibility
• /
• v.20 no.3
• /
• pp.141-150
• /
• 2017

The purpose of this study was to analyze the effect of the shape and attached position of E-textile-based stretchable sensors on motion-sensing performance and to investigate the requirements for the optimal structure of clothes for sensing limb motions. An experimental garment was prepared with different sensor shapes, and attachment positions. A child subject, wearing the experimental garment, performed arm and leg bending and extension motions at $60^{\circ}$, $90^{\circ}$ and $120^{\circ}$ motion angles, at a rate of 60 deg/sec. The changes in voltage triggered by the stretching and contracting of the fabric-sensor were measured, and an acceleration sensor was utilized to verify that the experimental motions were correctly performed. Dummy arms and legs of a child were manufactured to perform an identical test, in order to compare the dummy results with the actual human body experiment results. The analysis showed that the reproducibility and reliability of the rectangular sensor, showing uniform and stable were higher than those of the boat-shaped sensor, in both the dummy and the human body experiments. The attachment position of the sensor was more reproducible and reliable when placed on 4 cm below the elbow and knee joints in the dummy test, when placed in the joints of the elbow and knee, in children experiment. The appropriate shapes and attached positions of the sensor for sensing the motions were analyzed, and the results proved that motion-sensing of the human body is possible by utilizing flexible fabric-sensors integrated into clothes.