• Proceedings of the Korean Society of Machine Tool Engineers Conference
• /
• 2001.04a
• /
• pp.239-243
• /
• 2001

Neural networks are used in the framework of sensorbased tracking control of robot manipulators. They learn by practice movements the relationship between PSD (an analog Position Sensitive Detector) sensor readings for target positions and the joint commands to reach them. Using this configuration, the system can track or follow a moving or stationary object in real time. Furthermore, an efficient neural network architecture has been developed for real time learning. This network uses multiple sets of simple backpropagation networks one of which is selected according to which division (corresponding to a cluster of the self-organizing feature map) in data space the current input data belongs to. This lends itself to a very training and processing implementation required for real time control.

• Proceedings of the Korean Society of Machine Tool Engineers Conference
• /
• 2000.04a
• /
• pp.171-175
• /
• 2000

Neural networks are in the framework of sensorbased tracking control of robot manipulators. They learn by practice movements the relationship between PSD (an analog Position Sensitive Detector) sensor readings for target positions and the joint commands to reach them. Using this configuration, the system can track or follow a moving or stationary object in real time. Furthermore, an efficient neural network architecture has been developed for real time learning. This network uses multiple sets of simple backpropagation networks one of which is selected according to which division (corresponding to a cluster of the self-organizing feature map) in data space the current input data belongs to. This lends itself to a very training and processing implementation required for real time control.

• Proceedings of the Korean Society of Machine Tool Engineers Conference
• /
• 2000.10a
• /
• pp.108-112
• /
• 2000

Neural networks are used in the framework of sensorbased tracking control of robot manipulators. They learn by practice movements the relationship between PSD (an analog Position Sensitive Detector) sensor readings for target positions and the joint commands to reach them. Using this configuration, the system can track or follow a moving or stationary object in real time. Furthermore, an efficient neural network architecture has been developed for real time learning. This network uses multiple sets of simple backpropagation networks one of which is selected according to which division (corresponding to a cluster of the self-organizing feature map) in data space the current input data belongs to. This lends itself to a very training and processing implementation required for real time control.

• Journal of the Korean Society of Industry Convergence
• /
• v.7 no.2
• /
• pp.167-172
• /
• 2004

Neural networks are used in the framework of sensor based tracking control of robot manipulators. They learn by practice movements the relationship between PSD(an analog Position Sensitive Detector) sensor readings for target positions and the joint commands to reach them. Using this configuration, the system can track or follow a moving or stationary object in real time. Furthermore, an efficient neural network architecture has been developed for real time learning. This network uses multiple sets of simple back propagation networks one of which is selected according to which division (Corresponding to a cluster of the self-organizing feature map) in data space the current input data belongs to. This lends itself to a very training and processing implementation required for real time control.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.50 no.6
• /
• pp.201-211
• /
• 2013

In this paper, we propose a hand tracking and gesture recognition system. Our system employs a depth capture device to obtain 3D geometric information of user's bare hand. In particular, we build a flexible tracking volume and restrict the hand tracking area, so that we can avoid diverse problems caused by conventional object detection/tracking systems. The proposed system computes running average of the hand position, and tracking volume is actively adjusted according to the statistical information that is computed on the basis of uncertainty of the user's hand motion in the 3D space. Once the position of user's hand is obtained, then the system attempts to detect stretched fingers to recognize finger gesture of the user's hand. In order to test the proposed framework, we built a NUI system using the proposed technique, and verified that our system presents very stable performance even in the case that multiple objects exist simultaneously in the crowded environment, as well as in the situation that the scene is occluded temporarily. We also verified that our system ensures running speed of 24-30 frames per second throughout the experiments.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.36 no.5
• /
• pp.448-454
• /
• 2008

This paper deals with a behavior based decentralized control strategy for UAV swarming utilizing the artificial potential functions and the sliding mode control technique. Individual interactions for swarming behavior are modeled using the artificial potential functions. The motion of individual UAV is directed toward the negative gradient of the combined potential. For tracking the reference trajectory of UAV swarming, a swarming center is considered as the object of control. The sliding-mode control technique is adopted to make the proposed swarm control strategy robust with respect to the system uncertainties and the varying mission environment. Numerical simulation is performed to verify the performance of the proposed controller.

• Korean Institute of Interior Design Journal
• /
• v.22 no.6
• /
• pp.108-118
• /
• 2013

This study has analyzed the process of space information perceived and recognized through the estimation of observation frequency and number according to the time range of observation data acquired from observation experiment with the object of hospital lobby. The followings are the results analyzed at this study. First, the continual observation of 3 and 6 times was attentive and conscious for probing to find an object rather than for acquiring exact information and that of 9 times could be regarded as the time for acquiring visual appreciation. However, the repetitive occurrence of high and low frequencies can be thought of repetitive acts for visual appreciation. Second, the continual observation of 3 and 6 times had the highest observation frequency of II, while that of 9 times had the highest observation frequency of III. In case of 3 and 6 times, the observation frequency had the tendency to become a little higher after being low since V, and in case of 9 times it had the repetition of becoming low and high and from IX it characteristically got higher. This feature can be thought to be the process that the subject repeats the fixation and movement of observation at a visual activity for perception and recognition. In the process of first observation, the observation frequency was the highest after 20 seconds or so, but since then, it gets lower and repeatedly gets higher and lower as time passes. After 90 seconds, the frequency showed the tendency of getting higher continuously. Third, the examination of changing features of frequency may show the characteristics of exploration for and attention to space but if the observation frequency is not associated with observation times for analysis there will a limitation that the features of observation frequency cannot be clarified. Accordingly, the simultaneous analysis of both is very effective for estimating the observation characteristics seen at the processes of perception and recognition. Fourth, the general analysis of the both revealed: with the progress of observation time the discontinuous space exploration decreased, and as the observation time got longer the fixed attention to a specific spot increased. Fifth, in order to estimate the observation characteristics by the change of time range the observation frequency and times by trend line was analyzed, which approach seems to be an appropriate technique that can comprehensively show the overall flow of time series data.

• Science of Emotion and Sensibility
• /
• v.17 no.2
• /
• pp.111-122
• /
• 2014

This study has analyzed the scanning path algorithm shown at the process of exploring spatial information through an observation experiment with the object of lobby in subway station. In the estimation of observation time by section, the frequency of scanning type was found to increase as the observation time got longer, which makes it possible that the longer the observation lasts the more the observation interruptions occur. In addition, the observation slipped out of the range of imaging when any fatigue was caused from the observation or the more active exploration took place. Furthermore, when the trend line was employed for the examination of the changes to the scanning type by time section, "concentration" "diagonal or vertical" showed a sharp and a gentle increases along with the increase of time section respectively, while "circulation. combination, horizontal" showed a reduction. The observation data of the subjects observing a space include various visual information. The analysis of the scanning type found at "attention concentration" enabled to draw this significant conclusion. The features of increase and decrease of scanning types can be a fundamental data for understanding the scanning tendency by time.

• 한국HCI학회:학술대회논문집
• /
• 2008.02a
• /
• pp.829-834
• /
• 2008

Augmented reality makes the relationship between real world and virtual world. The basic goal of an AR system is to enhance the user's perception of and interaction with the real world through supplementing the real world with 3D virtual objects that appear to coexist in the same space as the real world. ARToolkit is widely used toolkit to develop any AR application. However, to do so, a lot of researcher faced many challenges and limitations. In this paper we comprehensively review some selected challenging issues using ARToolkit to develop AR system. And then, we implement a real application for augmented text book with sounds and 3D virtual objects which we called here AR Traffic Book concerning these challenging issues by ARToolkit. In this paper our foremost approaches on the most common challenging issues such as virtual object rendering, camera calibration and tracking.

• Journal of Internet Computing and Services
• /
• v.21 no.3
• /
• pp.113-121
• /
• 2020

Human Pose Estimation (HPE) which localizes the human body joints becomes a high potential for high-level applications in the field of computer vision. The main challenges of HPE in real-time are occlusion, illumination change and diversity of pose appearance. The single RGB image is fed into HPE framework in order to reduce the computation cost by using depth-independent device such as a common camera, webcam, or phone cam. However, HPE based on the single RGB is not able to solve the above challenges due to inherent characteristics of color or texture. On the other hand, depth information which is fed into HPE framework and detects the human body parts in 3D coordinates can be usefully used to solve the above challenges. However, the depth information-based HPE requires the depth-dependent device which has space constraint and is cost consuming. Especially, the result of depth information-based HPE is less reliable due to the requirement of pose initialization and less stabilization of frame tracking. Therefore, this paper proposes a new method of HPE which is robust in estimating self-occlusion. There are many human parts which can be occluded by other body parts. However, this paper focuses only on head self-occlusion. The new method is a combination of the RGB image-based HPE framework and the depth information-based HPE framework. We evaluated the performance of the proposed method by COCO Object Keypoint Similarity library. By taking an advantage of RGB image-based HPE method and depth information-based HPE method, our HPE method based on RGB-D achieved the mAP of 0.903 and mAR of 0.938. It proved that our method outperforms the RGB-based HPE and the depth-based HPE.