• Journal of Advanced Navigation Technology
• /
• v.28 no.1
• /
• pp.149-154
• /
• 2024

Research on particulate matter is advancing in real-time, and various methods are being studied to improve the accuracy of prediction models. Furthermore, studies that take into account various factors to understand the precise causes and impacts of particulate matter are actively being pursued. This paper trains an LSTM model using seasonal data and another LSTM model using concentration-based data. It compares and analyzes the PM_2.5 prediction performance of the two models. To train the model, weather data and air pollutant data were collected. The collected data was then used to confirm the correlation with PM_2.5. Based on the results of the correlation analysis, the data was structured for training and evaluation. The seasonal prediction model and the concentration-specific prediction model were designed using the LSTM algorithm. The performance of the prediction model was evaluated using accuracy, RMSE, and MAPE. As a result of the performance evaluation, the prediction model learned by concentration had an accuracy of 91.02% in the "bad" range of AQI. And overall, it performed better than the prediction model trained by season.

• Proceedings of the Korean Society of Near Infrared Spectroscopy Conference
• /
• 2001.06a
• /
• pp.1032-1032
• /
• 2001

Artificial Neural Network (ANN) calibration techniques have been used commercially for agricultural applications since the mid-nineties. Global models, based on transmission data from 850 to 1050 nm, are used routinely to measure protein and moisture in wheat and barley and also moisture in triticale, rye, and oats. These models are currently used commercially in approx. 15 countries throughout the world. Results concerning earlier European ANN models are being published elsewhere. Some of the findings from that study will be discussed here. ANN models have also been developed for coarsely ground samples of compound feed and feed ingredients, again measured in transmission mode from 850 to 1050 nm. The performance of models for pig- and poultry feed will be discussed briefly. These models were developed from a very large data set (more than 20,000 records), and cover a very broad range of finished products. The prediction curves are linear over the entire range for protein, fat moisture, fibre, and starch (measured only on poultry feed), and accuracy is in line with the performance of smaller models based on Partial Least Squares (PLS). A simple bias adjustment is sufficient for calibration transfer across instruments. Recently, we have investigated the possible use of ANN for a different type of NIR spectrometer, based on reflectance data from 1100 to 2500 nm. In one study, based on data for protein, fat, and moisture measured on unground compound feed samples, dedicated ANN models for specific product classes (cattle feed, pig feed, broiler feed, and layers feed) gave moderately better Standard Errors of Prediction (SEP) compared to modified PLS (MPLS). However, if the four product classes were combined into one general calibration model, the performance of the ANN model deteriorated only slightly compared to the class-specific models, while the SEP values for the MPLS predictions doubled. Brix value in molasses is a measure of sugar content. Even with a huge dataset, PLS models were not sufficiently accurate for commercial use. In contrast an ANN model based on the same data improved the accuracy considerably and straightened out non-linearity in the prediction plot. The work of Mr. David Funk (GIPSA, U. S. Department of Agriculture) who has studied the influence of various types of spectral distortions on ANN- and PLS models, thereby providing comparative information on the robustness of these models towards instrument differences, will be discussed. This study was based on data from different classes of North American wheat measured in transmission from 850 to 1050 nm. The distortions studied included the effect of absorbance offset pathlength variation, presence of stray light bandwidth, and wavelength stretch and offset (either individually or combined). It was shown that a global ANN model was much less sensitive to most perturbations than class-specific GIPSA PLS calibrations. It is concluded that ANN models based on large data sets offer substantial advantages over PLS models with respect to accuracy, range of materials that can be handled by a single calibration, stability, transferability, and sensitivity to perturbations.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.47 no.12
• /
• pp.831-840
• /
• 2019

Flush Air Data Sensing system (FADS) estimates air data states using pressure data measured at the surface of flight vehicles. The FADS system does not require intrusive probes, so it is suitable for high performance aircrafts, stealth vehicles, and hypersonic flight vehicles. In this study, calibration procedures and solution algorithms of the FADS for a sphere-cone shape vehicle are presented for the prediction of air data from subsonic to supersonic flights. Five flush pressure ports are arranged on the surface of nose section in order to measure surface pressure data. The algorithm selects the concept of separation for the prediction of flow angles and the prediction of pressure related variables, and it uses the pressure model which combines the potential flow solution for a subsonic flow with the modified Newtonian flow theory for a hypersonic flow. The CFD code which solves Euler equations is developed and used for the construction of calibration pressure data in the Mach number range of 0.5~3.0. Tests are conducted with various flight conditions for flight Mach numbers in the range of 0.6~3.0 and flow angles in the range of -10°~+10°. Air data such as angle of attack, angle of sideslip, Mach number, and freestream static pressure are predicted and their accuracies are analyzed by comparing predicted data with reference data.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.48 no.2
• /
• pp.106-116
• /
• 2011

USN(Ubiquitous Sensor Network) environment permits us to access whatever information we want, whenever we want. The technologies to provide a basement to these environments premise an accurate location establishment. Especially, ITS(Intelligent Transportation Systems) is easily constructed by applying USN technology. Localization can be categorized as either Range-based or Range-free. Range-based is known to be not suitable for the localization based on sensor network, because of the irregularity of radio propagation and the additional device requirement. The other side, Range-free is much appropriated for the resource constrained sensor network because it can actively locate by means of the communication radio. But, generally the location accuracy of Range-free is low. Especially, it is very low in a low-density environment. So, these two methods have both merits and demerits. Therefore, it requires a new method to be able to improve tracking accuracy by combining the two methods. This paper proposes the tracking scheme based on range-hybrid, which can markedly enhance tracking accuracy by effectively using the information of surrounding nodes and the RSSI(Received Signal Strength Indication) that does not require additional hardware. Additionally, we present a method, which can improve the accuracy of vehicle tracking by adopting the prediction mechanism. Simulation results show that our method outperforms other methods in the transportation simulation environment.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.47 no.2
• /
• pp.54-64
• /
• 2010

Localization is the most important feature in the sensor network environment because it is a basic element enabling people and things to aware the circumference environment. Existing localization methods can be categorized as either range-based or range-free. While range-based is known to be not suitable because of the irregularity of radio propagation and the additional device requirement. range-free is much appropriated for the resource constrained sensor network because it can actively locate by means of the communication radio. But its location accuracy is just depended on the density of circumference nodes; it is very low in low-density sensor network environment. This paper proposes a mobile object tracking method, named DRTS(Distributed Range-hybrid Tracking Scheme), with combining range-based and range-free. It is optimally making use of the location, communication range, and received signal strength from circumference nodes. Especially, it can greatly improve the mobile tracking accuracy by adapting a new prediction method, named EGP(Estimative Gird Points) into the proposed location estimation method. The simulation results show that our method outperforms the other localization and tracking methods in the tracking accuracy point of view.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.16 no.12
• /
• pp.8222-8227
• /
• 2015

The visibility range is defined from where one can see, which can be changed by the character sizes and illuminances and so on, which of one-hundred and twelve students are measured for three illuminances and three character sizes in this paper. In determining the character sizes and illuminances, the visibility range can be an important data. Functions are proposed whose independent variable is illuminance and whose dependent variable is visibility range in order to predict the visibility range of unmeasured illuminances. The fractional functions are used for three character sizes because the visibility range is invariant according to illuminance. There are three parameters to be determined - k, m, n, which are selected based on the measured visibility ranges. Because the visibility ranges of three character sizes are measured, three k's can be calculated. In this paper the case of minimum variance of three k's is selected, and three parameters - k,m,n- of that case is selected. The three functions according to three character sizes are proposed. The small differences between the measured data and the postulated functions verifies the accuracy of the functions.

• Atmosphere
• /
• v.34 no.1
• /
• pp.23-34
• /
• 2024

The Korea Meteorological Administration has improved medium-range weather forecasts by implementing post-processing methods to minimize numerical model errors. In this study, we employ a statistical correction technique known as the minimum continuous ranked probability score (CRPS) to refine medium-range forecast guidance. This technique quantifies the similarity between the predicted values and the observed cumulative distribution function of the Unified Model Ensemble Prediction System for Global (UM EPSG). We evaluated the performance of the medium-range forecast guidance for surface air temperature and relative humidity, noting significant enhancements in seasonal bias and root mean squared error compared to observations. Notably, compared to the existing the medium-range forecast guidance, temperature forecasts exhibit 17.5% improvement in summer and 21.5% improvement in winter. Humidity forecasts also show 12% improvement in summer and 23% improvement in winter. The results indicate that utilizing the minimum CRPS for medium-range forecast guidance provide more reliable and improved performance than UM EPSG.

• Asian-Australasian Journal of Animal Sciences
• /
• v.25 no.12
• /
• pp.1748-1758
• /
• 2012

Two experiments were conducted to evaluate the effects of chemical composition of wheat shorts and red dog on energy and amino acid digestibility in growing pigs and to establish prediction models to estimate their digestible (DE) and metabolizable (ME) energy content and as well as their standardized ileal digestible (SID) amino acid content. For Exp. 1, sixteen diets were fed to thirty-two growing pigs according to a completely randomized design during three successive periods. The basal diet was based on corn and soybean meal while the other fifteen diets contained 28.8% wheat shorts (N = 7) or red dog (N = 8), added at the expense of corn and soybean meal. Over the three periods, each diet was fed to six pigs with each diet being fed to two pigs during each period. The apparent total tract digestibility (ATTD) of energy in wheat shorts and red dog averaged 75.1 and 87.9%. The DE values of wheat shorts and red dog averaged 13.8 MJ/kg (range 13.1 to 15.0 MJ/kg) and 15.1 MJ/kg (range 13.3 to 16.6 MJ/kg) of dry matter, respectively. For Exp. 2, twelve growing pigs were allotted to two $6{\times}6$ Latin Square Designs with six periods. Ten of the diets were formulated based on 60% wheat shorts or red dog and the remaining two diets were nitrogen-free diets based on cornstarch and sucrose. Chromic oxide (0.3%) was used as an indigestible marker in all diets. There were no differences (p>0.05) in SID values for the amino acids in wheat shorts and red dog except for lysine and methionine. Apparent ileal digestibility (AID) and SID values for lysine in different sources of wheat shorts or red dog, which averaged 78.1 and 87.8%, showed more variation than either methionine or tryptophan. A stepwise regression was performed to establish DE, ME and amino acid digestibility prediction models. Data indicated that fiber content and amino acid concentrations were good indicators to predict energy values and amino acid digestibility, respectively. The present study confirms the large variation in the energy content and amino acid digestibility in wheat shorts and red dog, and describes the factors that influence this variation and presents equations based on chemical composition that could probably be used to predict the DE and ME values as well as the amino acid digestibility of wheat shorts and red dog.

• Fire Science and Engineering
• /
• v.33 no.6
• /
• pp.20-27
• /
• 2019

In this study, a new function with higher accuracy for fire heat release rate prediction was developed. The 'αt²' curve, which is the major exponential function currently used for fire engineering calculations, must be improved to minimize the prediction gap that causes fire system engineering inefficiency and lower cost-effectiveness. The newly developed prediction function was designed to cover the initial fire stage that features rapid growth based on logistic function theory, which has a more logical background and graphical similarity compared to conventional exponential function methods for 'αt²'. The new function developed in this study showed apparently higher prediction accuracy over wider range of fire growth durations. With the progress of fire growth pattern studies, the results presented herein will contribute towards more effective fire protection engineering.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.13 no.1
• /
• pp.101-107
• /
• 2013

Is generated in real time in the real world, a large amount of time series data from a wide range of business areas. But it is not easy to determine the optimal model for the description and understanding of the time series data is represented as a dynamic feature. In this study, through the HMM suitable for estimating the short and long-term forecasting model of time-series data to estimate a model that can explain the characteristics of these time series data, it was estimated to predict future patterns of movement. The actual stock market through various materials, information criterion and optimal model estimation for the length of the most efficient data was found to accurately estimate the state of the model. Similar movement patterns predictive than the long-term prediction is more similar to the short-term prediction of the experimental result were found to be.