• Journal of Institute of Control, Robotics and Systems
• /
• v.22 no.9
• /
• pp.773-777
• /
• 2016

In this paper, a novel multi-object tracking method to track an unknown number of objects is proposed. To handle multiple object states and uncertain observations efficiently, a probability hypothesis density (PHD) filter is adopted and modified. The PHD filter is capable of reducing false positives, managing object appearances and disappearances, and estimating the multiple object trajectories in a unified framework. Although the PHD filter is robust in cluttered environments, it is vulnerable to false negatives. For this reason, we propose to exploit local observations in an RFS of the observation model. Each local observation is generated by using an online trained object detector. The main purpose of the local observation is to deal with false negatives in the PHD filtering procedure. The experimental results demonstrated that the proposed method robustly tracked multiple objects under practical situations.

• 제어로봇시스템학회:학술대회논문집
• /
• 1991.10a
• /
• pp.476-481
• /
• 1991

This paper investigates the relationship between the two problems, supervisor reduction and observation function (projection) design, which arise in supervisory control of DEDS. It is shown through an example that a reduced supervisor of minimal size does not necessarily result in a maximal projection when a projection design method which uses the transition structure of a supervisor is applied. Also, if an L-realizable projection P is available and if a supervisor has a special structural feature, a cover C for supervisor reduction can be easily obtained. By investigating the control-compatibility of states of the reduced supervisor based on C, we can also check maximality of P in a simple manner.

• International Journal of Control, Automation, and Systems
• /
• v.6 no.3
• /
• pp.306-319
• /
• 2008

Reaching fault tolerance in technological systems requires to detect malfunctions. This paper presents a diagnostic method that is robust with respect to unknown-but-bounded uncertainties of the dynamical model and the measurements. By using models of the faultless and the faulty behaviours, a state-set observer computes polyhedral sets from which the consistency of the models with the interval measurements is determined. The diagnostic result is proven to be complete, i.e., the set of faults obtained by the diagnostic algorithm includes the actual fault. The algorithm is illustrated by an application example.

• Proceedings of the KIEE Conference
• /
• summer
• /
• pp.1278-1280
• /
• 2004

In this paper, we propose a new diagnosis method of DC/DC converter aging. The method is based on the variations of parasitic resistor for the aging process. We apply an on-line diagnosis of DC/DC converter because the observation is not a device, but a system. This study proposes a method of DC/DC converter diagnosis by analyzing the variations of model on the variations of parasitic resistor.

• The Journal of the Petrological Society of Korea
• /
• v.23 no.2
• /
• pp.119-138
• /
• 2014

There are various ways to monitor active volcanoes, such as the method of observing the activity of a volcano with the naked eye, the method of referring to the past eruptive history based on the historic records and the method of monitoring volcanoes by using observation equipment. The most basic method from the observation equipment-using methods to monitor volcanoes is seismic monitoring. In addition to this, the ways to monitor volcanoes are as follows: resonance observation which may be effective to remove artificial noises from the seismic activities that are recorded in the seismograph, ground deformation by using precision leveling, electronic distance measurement, tiltmeter, GPS, and InSAR observation method, volcanic gas monitoring, hydrologic and meteorological monitoring, and other geophysical monitoring methods. These monitoring methods can make volcanic activities effectively monitored, determine the behavior of magmas in magma chambers and help predict the future volcanic eruptions more accurately and early warning, thus, minimize and mitigate the damage of volcanic hazards.

• 제어로봇시스템학회:학술대회논문집
• /
• 1992.10b
• /
• pp.272-275
• /
• 1992

A novel fault diagnosis method based on likelihood decomposition is proposed for linear stochastic systems described by autoregressive (AR) model. Assuming that at some time instant .tau. the fault of one of the following two types is occurs: innovation fault (actuator fault); and observation fault (sensor fault), the log-likelihood function is decomposed into two components based on the observations before and after .tau., respectively, Then, the type of the fault is determined by comparing the log-likelihoods corresponding two types of faults. Numerical examples demonstrate the usefulness of the proposed diagnosis method.

• The Journal of the Acoustical Society of Korea
• /
• v.39 no.6
• /
• pp.615-621
• /
• 2020

In this paper, we propose a multi-site based earthquake event classification method using graph convolution networks. In the traditional earthquake event classification methods using deep learning, they used single-site observation to estimate seismic event class. However, to achieve robust and accurate earthquake event classification on the seismic observation network, the method using the information from the multi-site observations is needed, instead of using only single-site data. Firstly, our proposed model employs convolution neural networks to extract informative embedding features from the single-site observation. Secondly, graph convolution networks are used to integrate the features from several stations. To evaluate our model, we explore the model structure and the number of stations for ablation study. Finally, our multi-site based model outperforms up to 10 % accuracy and event recall rate compared to single-site based model.

• Atmosphere
• /
• v.28 no.4
• /
• pp.357-367
• /
• 2018

In this study, the temperature bias of aircraft observation was estimated through comparison with sonde observation prior to developing the temperature bias correction method at the Korea Institute of Atmospheric Prediction Systems (KIAPS). First, we tried to compare aircraft temperature with collocated sonde observations at 0000 UTC on June 22, 2012. However, it was difficult to estimate the temperature bias due to the lack of samples and the uncertainty of the sonde position at high altitudes. Second, we attempted a background innovation comparison for sonde and aircraft using KIAPS Package for Observation Processing (KPOP). The one month averaged background innovation shows the aircraft temperature have a warm bias against sonde for all levels. In particular, there is a globally distinct warm bias about 0.4 K between 200 hPa and 300 hPa corresponding to flight level. Spatially, most of the areas showed the warm bias except for below 300 hPa in some part of China at 0000 and 1200 UTC and below 850 hPa in Australia at 0000 UTC. In general, the temperature bias was larger at 1200 UTC than 0000 UTC. Based on the estimated temperature bias, we have applied the static bias correction method to the aircraft temperature observation. As a result, the warm bias of the aircraft temperature has decreased at most levels, but a slight cold bias has occurred in some areas.

• Journal of Animal Science and Technology
• /
• v.65 no.2
• /
• pp.441-458
• /
• 2023

The poultry industry, which produces excellent sources of protein, suffers enormous economic damage from diseases. To solve this problem, research is being conducted on the early detection of infection according to the behavioral characteristics of poultry. The purpose of this study was to evaluate the potential of a non-movement behavior observation method to detect sick chickens. Forty 1-day-old Ross 308 males were used in the experiments, and an isolator equipped with an Internet Protocol (IP) camera was fabricated for observation. The chickens were inoculated with Salmonella enterica serovar Gallinarum A18-GCVP-014, the causative agent of fowl typhoid (FT), at 14 days of age, which is a vulnerable period for FT infection. The chickens were continuously observed with an IP camera for 2 weeks after inoculation, chickens that did not move for more than 30 minutes were detected and marked according to the algorithm. FT infection was confirmed based on clinical symptoms, analysis of cardiac, spleen and liver lesion scores, pathogen re-isolation, and serological analysis. As a result, clinical symptoms were first observed four days after inoculation, and dead chickens were observed on day six. Eleven days after inoculation, the number of clinical symptoms gradually decreased, indicating a state of recovery. For lesion scores, dead chickens scored 3.57 and live chickens scored 2.38. Pathogens were re-isolated in 37 out of 40 chickens, and hemagglutination test was positive in seven out of 26 chickens. The IP camera applied with the algorithm detected about 83% of the chickens that died in advance through non-movement behavior observation. Therefore, observation of non-movement behavior is one of the ways to detect infected chickens in advance, and it appears to have potential for the development of remote broiler management system.

• Journal of Wetlands Research
• /
• v.25 no.1
• /
• pp.48-63
• /
• 2023

Research on the method of calculating and estimating evapotranspiration has been steadily conducted. Various models have been developed according to different backgrounds, and each of these models has different characteristics such as required input data. Therefore, this study introduces the theoretical background and characteristics of evapotranspiration models and the development process of domestic research on evapotranspiration by era. First, the origin and theoretical background of the potential evapotranspiration models are summarized in addition to classifying them by input data. Then, the characteristics of the actual evapotranspiration estimation methods are summarized. Additionally, methods based on observation and methods using the rainfall-runoff models are summarized.