• International conference on construction engineering and project management
• /
• 2022.06a
• /
• pp.814-822
• /
• 2022

The deployment of sensors for Structural Health Monitoring requires a complicated network arrangement, ground truthing, and calibration for validating sensor performance periodically. Any conventional sensor on a structural element is also subjected to static and dynamic vertical loadings in conjunction with other environmental factors, such as brightness, noise, temperature, and humidity. A structural model with strain gauges was built and tested to get realistic sensory information. This paper investigates different deep learning architectures and algorithms, including unsupervised, autoencoder, and supervised methods, to benchmark blind drift calibration methods using deep learning. It involves a fully connected neural network (FCNN), a long short-term memory (LSTM), and a gated recurrent unit (GRU) to address the blind drift calibration problem (i.e., performing calibrations of installed sensors when ground truth is not available). The results show that the supervised methods perform much better than unsupervised methods, such as an autoencoder, when ground truths are available. Furthermore, taking advantage of time-series information, the GRU model generates the most precise predictions to remove the drift overall.

• Proceedings of the Korean Statistical Society Conference
• /
• 2005.11a
• /
• pp.85-90
• /
• 2005

This paper proposes the calibration procedure for stratified Warner's randomized response model, which suggested by Kim and Warde (2004). It is shown that the proposed calibration estimator is more efficient than the Kim and Warde's model.

• Journal of Sensor Science and Technology
• /
• v.13 no.1
• /
• pp.27-34
• /
• 2004

Camera calibration is an important and fundamental procedure for the application of a vision sensor to 3D problems. Recently many camera calibration methods have been proposed particularly in the area of robot vision. However, the reliability of data used in calibration has been seldomly considered in spite of its importance. In addition, a camera model can not guarantee good results consistently in various conditions. This paper proposes methods to overcome such uncertainty problems of data and camera models as we often encounter them in practical camera calibration steps. By the use of the RANSAC (Random Sample Consensus) algorithm, few data having excessive magnitudes of errors are excluded. Artificial neural networks combined in a two-step structure are trained to compensate for the result by a calibration method of a particular model in a given condition. The proposed methods are useful because they can be employed additionally to most existing camera calibration techniques if needed. We applied them to a linear camera calibration method and could get improved results.

• Journal of the Korean Society for Precision Engineering
• /
• v.14 no.4
• /
• pp.64-71
• /
• 1997

Objective of calibration is to find out the accurate kinematic relationships between robot joint angles and the position of the end-effector by estimating accurate model parameters defining the kinematic function. Estimating the model parameters requires measurement of the end-effector position at a number of different robot configurations. This paper studies the implication of measurement configurations in robot calibration. For selecting appropriate measurement configurations in robot calibration, an index is defined to measure the observability of the model parameters with respect to a set of robot configurations. It is found that, as the observability index of the selected measurement configurations increase the attribution of the position errors to the parameter errors becomes dominant while the effects of the measurement and unmodeled errors are less significant; consequently better estimation of parameter errors is expected. To demonstrate the implication of the observability measure in robot calibration, computer simulations are performed and their results are discussed.

• ETRI Journal
• /
• v.26 no.6
• /
• pp.605-614
• /
• 2004

In this paper, we investigate array calibration algorithms to derive a further improved version for correcting antenna array errors and RF transceiver errors in CDMA smart antenna systems. The structure of a multi-channel RF transceiver with a digital calibration apparatus and its calibration techniques are presented, where we propose a new RF receiver calibration scheme to minimize interference of the calibration signal on the user signals. The calibration signal is injected into a multi-channel receiver through a calibration signal injector whose array response vector is controlled in order to have a low correlation with the antenna response vector of the receive signals. We suggest a model-based antenna array calibration to remove the antenna array errors including mutual coupling errors or to predict the element patterns from the array manifold measured at a small number of angles. Computer simulations and experiment results are shown to verify the calibration algorithms.

• Journal of Korean Society for Quality Management
• /
• v.34 no.2
• /
• pp.122-128
• /
• 2006

One feature of a computer simulation experiment, different from a physical experiment, is that the output is often deterministic. Moreover the codes are computationally very expensive to run. This paper deals with the design and analysis of computer experiments(DACE) which is a relatively new statistical research area. We model the response of computer experiments as the realization of a stochastic process. This approach is basically the same as using a spatial linear model. Applications to the optimal mechanical designing and model calibration problems are illustrated. Algorithms for selecting the best spatial linear model are also proposed.

• KIEE International Transactions on Electrophysics and Applications
• /
• v.5C no.1
• /
• pp.14-17
• /
• 2005

In this paper, an automated methodology for generating model parameters considering real manufacturing processes is presented with verified results. In addition, the outcomes of applications to the next generation of flash memory devices using the parameters calibrated from the process specification decision are analyzed. The test vehicle is replaced with a well-calibrated TCAD simulation. First, the calibration methodology is introduced and tested for a flash memory device. The calibration errors are less than 5% of a full chip operation, which is acceptable to designers. The results of the calibration are then used to predict the I-V curves and the model parameters of various transistors for the design of flash devices.

• Transactions on Electrical and Electronic Materials
• /
• v.5 no.5
• /
• pp.180-184
• /
• 2004

This paper proposes a novel methodology of systematic global calibration and validates its accuracy and efficiency with application to memory and logic devices. With 175 SIMS profiles which cover the range of conditions of implant and diffusion processes in the fabrication lines, the dominant diffusion phenomenon in each process temperature region has been determined. Using the dual-pearson implant model and fully-coupled diffusion model, the calibration was performed systematically. We applied the globally calibrated process simulator parameters to memory and logic devices to predict the optimum process conditions for target device characteristics.

• Journal of the Korean Society for Precision Engineering
• /
• v.13 no.1
• /
• pp.131-142
• /
• 1996

Most industrial robots cannot be off-line programmed to carry out a task accurately, unless their kinematic model is suitably corrected through a calibration procedure. However, normal calibration is an expensive and time-consumming precedure due to the highly accurate measurement equipment required and due to the significant amount of data that must be collected. This paper presents a simple and economic procedure to improve the efficiency of robot calibration especially for arc welding application. To simplify the measurement process, an automotic data measurement algorithm as well as a simple measurement device are developed. Also, a calibration algorithm which can automatically identify the independent model parameters to be estimated is presented. To demonstrated the simplicity and the effectiveness of the procedure, experimental studies and computer simulations are performed and their results are discussed.

• Communications for Statistical Applications and Methods
• /
• v.15 no.4
• /
• pp.597-603
• /
• 2008

In this paper, we propose the calibration procedure for the valiance reduction of the stratified Warner's randomized response estimators, which suggested by Hong et al. (1994) and Kim and Warde (2004), using auxiliary information at the population level. It is shown that the proposed calibration estimators are more efficient than the ordinary Warner's estimators.