• Journal of the Institute of Convergence Signal Processing
• /
• v.6 no.2
• /
• pp.82-88
• /
• 2005

A compensation scheme for a class of linearly parameterized systems is presented. The compensator consists of a typical linearizing control and an adaptive observer with integration type update law, which is based on Speed Gradient (SG) algorithm.. Instead of the intermediate functions of the compensation schemes suggested by other researchers, the proposed compensator is designed with some design functions which guarantee the growth, convexity, attainability, and pseudo gradient conditions in the update law. The scheme achieves the asymptotic stability of the tracking error and the boundedness of the estimation errors. A numerical example is given to demonstrate the validity of the proposed design.

• International journal of advanced smart convergence
• /
• v.8 no.4
• /
• pp.68-74
• /
• 2019

Along with the deeper architecture in the deep learning approaches, the need for the data becomes very big. In the real problem, to get huge data in some disciplines is very costly. Therefore, learning on limited data in the recent years turns to be a very appealing area. Meta-learning offers a new perspective to learn a model with this limitation. A state-of-the-art model that is made using a meta-learning framework, Meta-SGD, is proposed with a key idea of learning a hyperparameter or a learning rate of the fast adaptation stage in the outer update. However, this learning rate usually is set to be very small. In consequence, the objective function of SGD will give a little improvement to our weight parameters. In other words, the prior is being a key value of getting a good adaptation. As a goal of meta-learning approaches, learning using a single gradient step in the inner update may lead to a bad performance. Especially if the prior that we use is far from the expected one, or it works in the opposite way that it is very effective to adapt the model. By this reason, we propose to add a weight term to decrease, or increase in some conditions, the effect of this prior. The experiment on few-shot learning shows that emphasizing or weakening the prior can give better performance than using its original value.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2003.11a
• /
• pp.840-840
• /
• 2003

Operational Modal Analysis (OMA) is a technique for identification of modal parameters by measurement of only the system's response. On many lightweight structures, such as load-speaker cones and disk drive read/write heads, is impossible or impractical to measure the input forces. Another characteristic of lightweight structure is their sensitivity to mass loading from sensors. The Scanning Laser Doppler Vibrometry(SLDV) allows response measurements to be taken without mass loading. One disadvantage of OMA testing compared to tradition input output modal testing is the OMA mode shapes are un-scaled. This means that the mode shape obtained from an OMA test can not used for analytical structural modification studies. However, the un-scaled mode shapes from an OMA test can be used to update a Finite Element Model (FEM). The updated FEM can then be used to analytically predict the effect of structural modifications. This paper will present the results of an OMA test performed on a simple plate and motor in operating conditions. The un-scaled mode shapes from this test will be used to update a FEM model of the system. The updated FEM model will be then be used to predict the effect of attaching a mass to the plate. The shapes predicted by the FEM for the modified system will be compared to a second OMA test on the modified system

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.11 no.2
• /
• pp.118-123
• /
• 2011

Video based object tracking normally deals with non-stationary image streams that change over time. Robust and real time moving object tracking is considered to be a problematic issue in computer vision. Multiple object tracking has many practical applications in scene analysis for automated surveillance. In this paper, we introduce a specified object tracking based particle filter used in an environment of multiple moving objects. A differential image region based tracking method for the detection of multiple moving objects is used. In order to ensure accurate object detection in an unconstrained environment, a background image update method is used. In addition, there exist problems in tracking a particular object through a video sequence, which cannot rely only on image processing techniques. For this, a probabilistic framework is used. Our proposed particle filter has been proved to be robust in dealing with nonlinear and non-Gaussian problems. The particle filter provides a robust object tracking framework under ambiguity conditions and greatly improves the estimation accuracy for complicated tracking problems.

• Smart Structures and Systems
• /
• v.15 no.3
• /
• pp.751-767
• /
• 2015

A new procedure is proposed for assessing probabilistic condition of structures considering effect of measured data uncertainty. In this procedure, multiple Finite Element (FE) models are identified by using weighting vectors that represent the uncertainty conditions of measured data. The distribution of structural parameters is analysed using a Principal Component Analysis (PCA) in relation to uncertainty conditions, and the identified models are classified into groups according to their similarity by using a K-means method. The condition of a structure is then assessed probabilistically using FE models in the classified groups, each of which represents specific uncertainty condition of measured data. Yeondae bridge, a steel-box girder expressway bridge in Korea, is used as an illustrative example. Probabilistic condition of the bridge is evaluated by the distribution of load rating factors obtained using multiple FE models. The numerical example shows that the proposed method can quantify uncertainty of measured data and subsequently evaluate efficiently the probabilistic condition of bridges.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2024.05a
• /
• pp.794-797
• /
• 2024

Despite advancements in deep learning, existing semantic segmentation models exhibit suboptimal performance under adverse weather conditions, such as fog or rain, whereas they perform well in clear weather conditions. To address this issue, much of the research has focused on making image or feature-level representations weather-independent. However, disentangling the style and content of images remains a challenge. In this work, we propose a novel fine-tuning method, 'freeze-n-update.' We identify a subset of model parameters that are weather-independent and demonstrate that by freezing these parameters and fine-tuning others, segmentation performance can be significantly improved. Experiments on a test dataset confirm both the effectiveness and practicality of our approach.

• Journal of Navigation and Port Research
• /
• v.27 no.4
• /
• pp.425-430
• /
• 2003

The satellite navigation system is widely used for identifying a user's position regardless of weather or geographic conditions and also make effect on new technology of marine LBS(Location Based Service). which has the technology of geographic information such as the ENC. Generally, there are conceivable systems of marine LBS such as ECDIS, or ECS that use the ENC itself with powerful processor in installed type on ships bridge. Since the ENC is relatively heavy structure with dummy format for data transfer between different systems, we should reduce the ENC to small and compact size in order to use it in mobile platform. In this paper, we assumed that the mobile system like PDA, or Webpad can be used for small navigation or information system in marine field. We considered the reduction of the ENC size to make them fit well to small capability of mobile platform. However, the ENC should be updated periodically by update profile data produced by HO. If we would reduce the ENC without a consideration of update, we could not get newly updated data furthermore. As summary, we studied considerations for ENC reduction with update capability. It will make the ENC be useful in many low performance platforms for various applications.

• Journal of Sensor Science and Technology
• /
• v.28 no.6
• /
• pp.345-354
• /
• 2019

The 3D position of pedestrians is a physical quantity used in various fields, such as automotive navigation and augmented reality. An inertial navigation system (INS) based pedestrian dead reckoning (PDR), hereafter INS-PDR, estimates the relative position of pedestrians using an inertial measurement unit (IMU). Since an INS-PDR integrates the accelerometer signal twice, cumulative errors occur and cause a rapid increase in drifts. Various correction methods have been proposed to reduce drifts. For example, one of the most commonly applied correction method is the zero velocity update (ZUPT). This study investigated the characteristics of the existing INS-PDR methods based on shoe-mounted IMU and compared the estimation performances under various conditions. Four methods were chosen: (i) altitude correction (AC); (ii) step length correction (SLC); (iii) advanced heuristic drift elimination (AHDE); and (iv) magnetometer-based heading correction (MHC). Experimental results reveal that each of the correction methods shows condition-sensitive performance, that is, each method performs better under the test conditions for which the method was developed than it does under other conditions. Nevertheless, AC and AHDE performed better than the SLC and MHC overall. The AC and AHDE methods were complementary to each other, and a combination of the two methods yields better estimation performance.

• International Journal of Naval Architecture and Ocean Engineering
• /
• v.11 no.1
• /
• pp.233-243
• /
• 2019

In this paper, numerical and experimental studies are performed to investigate the liquid impact on a free falling wedge. In the numerical simulation, the structure is assumed to be rigid and the elastic response is ignored. The fully nonlinear coupling between wedge and water is considered by an auxiliary function method based on the Boundary Element Method (BEM). At the intersection of the wedge surface and liquid surface, two coincident nodes are used to decouple the boundary conditions. The Eulerian free surface conditions in the local coordinate system are adopted to update the deformed free surface. In the experiments, five pressure sensors are fixed on each side of the wedge which is released from an experimental installation. Steel and aluminum wedges that have different structural elasticity are used in the experiments to investigate the influence of structural elasticity on the impact force. Numerical results are compared with experimental data and they agree very well. The influence of fluid gravity, body mass, initial entry speed and deadrise angle on the impact pressure are further investigated.

• Nuclear Engineering and Technology
• /
• v.53 no.11
• /
• pp.3723-3740
• /
• 2021

The harsh conditions in the reactor affect the thermal and mechanical performance of the fuel plate heavily. Some in-pile behaviors, like fission-induced swelling, can cause a large deformation of fuel plate at very high burnup, which may even disturb the flow of coolant. In this research, the emphasis is put on the thermal expansion, fission-induced swelling, interaction layer (IL) growth, creep of the fuel meat, and plasticity of the cladding for the U₃Si₂/Al dispersion fuel plate. A detailed model of the fuel meat swelling is developed. Taking these in-pile behaviors into consideration, the three-dimensional large deformation incremental constitutive relations and stress update algorithms have been developed to study its thermal-mechanical performance under normal conditions using Abaqus. Results have shown that IL can effectively decrease the thermal conductivity of fuel meat. The high Mises stress region mainly locates at the interface between fuel meat and cladding, especially around the side edge of the interface. With irradiation time increasing, the stress in the fuel plate gets larger resulting from the growth of fuel meat swelling but then decreases under the effect of creep deformation. For the cladding, plasticity deformation does not occur within the irradiation time.