• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.16 no.5
• /
• pp.331-337
• /
• 2023

In this paper, we propose a deep learning model to detect lumbar 3 (L3) CT images to determine the occurrence and degree of sarcopenia. In addition, we would like to propose an optimization technique that uses oversampling ratio and class weight as design parameters to address the problem of performance degradation due to data imbalance between L3 level and non-L3 level portions of CT data. In order to train and test the model, a total of 150 whole-body CT images of 104 prostate cancer patients and 46 bladder cancer patients who visited Gangneung Asan Medical Center were used. The deep learning model used ResNet50, and the design parameters of the optimization technique were selected as six types of model hyperparameters, data augmentation ratio, and class weight. It was confirmed that the proposed optimization-based L3 level extraction model reduced the median L3 error by about 1.0 slices compared to the control model (a model that optimized only 5 types of hyperparameters). Through the results of this study, accurate L3 slice detection was possible, and additionally, we were able to present the possibility of effectively solving the data imbalance problem through oversampling through data augmentation and class weight adjustment.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.41 no.1
• /
• pp.70-78
• /
• 2013

Before GBAS ground systems is installed at the airport, the site survey is needed to determine the suitability of proposed GBAS candidate sites depending on the siting requirements. Therefore, analysis of GPS signal reception environment, one of the site survey steps, is required. In this paper, the number of visible satellites, GPS signal strength, multipath error, radio frequency interference and predicted availability were analyzed using the GPS data of Gimpo International Airport measured by PortaSAT equipments and the analysis results were represented.

• Journal of Navigation and Port Research
• /
• v.37 no.1
• /
• pp.41-48
• /
• 2013

A positioning technique using the augmentation system has been researched to improve the accuracy. The network RTK is the precise positioning technique using carrier phase correction data from reference stations and is constantly being researched. The study for the system accuracy has been performed but system integrity research has not been done as much as system accuracy. In this paper, we presented the anomaly detection algorithm by satellite system and the diagnosis algorithm to a basic research in the integrity on network RTK. And the presented algorithms are verified on the DL-V3 dual-frequency receiver and the simulated error scenario using the GSS7700.

• Journal of Advanced Navigation Technology
• /
• v.19 no.3
• /
• pp.207-216
• /
• 2015

Since the distribution of SBAS reference stations is one of the most important factors that affect the system performance, the effects of the distribution should be analyzed carefully from the beginning of the program to develop the system conforming to given performance requirements. The reference stations of KASS, the Korean SBAS, are planned to be installed only inside South Korea, which limits the number and area of those. It differentiates KASS from others that have much larger sites. In this paper, the author analyzes the performance impact on UDRE and ${\delta}UDRE$ for GPS and GEO due to the limitations by a series of simulations, which showed that the UDRE performance depends on the diversity of the reference station distribution and the impact on the GEO UDRE is significant. The paper concludes by providing KASS design and development considerations to minimize the possible performance risks due to the limitations of KASS reference station distribution.

• Journal of Institute of Control, Robotics and Systems
• /
• v.20 no.11
• /
• pp.1103-1111
• /
• 2014

The results of control law design for a tilt-rotor unmanned aerial vehicle that has a nacelle mounted wing extension (WE) are presented in this paper. It consists of a control surface mixer, stability and control augmentation system (SCAS), hold mode for altitude / speed / heading, and a guidance mode for preprogram and point navigation which includes automatic take-off and landing. The conversion corridor and the control moments derivatives between the original tilt-rotor and its variant of the nacelle mounted WE were compared to show the effectiveness of the WE. The nacelle conversion of the original tilt-rotor starts when the airspeed is greater than 30 km/h but its WE variant starts at 0 km/h in order to reduce the drag caused by the high incidence angle of the WE. The stability margins of the inner loop are presented with the optimization approach. The outer loops for the hold mode are designed with trial and error methods with linear and nonlinear simulation. The main control parameter for altitude control of the helicopter mode is thrust command and it is transferred to the pitch attitude command in airplane mode. Otherwise, the control parameter for the speed of the helicopter mode is the pitch attitude command and it is transferred to the thrust command in airplane mode. Therefore the speed and altitude hold mode are coupled to each other and are engaged at the same time when an internal pilot engages any of the altitude or speed hold modes. The nonlinear simulation results of the guidance control for the preprogrammed mode and point navigation are also presented including automatic take-off and landing in order to prove the full control law.

• Journal of Aerospace System Engineering
• /
• v.8 no.4
• /
• pp.24-31
• /
• 2014

The modern version of aircrafts is allowed to guarantee the superior handing qualities within the entire flight envelope by imposing the adequate stability and flying qualities on a target aircraft through the various techniques of flight control law design. Generally, the flight control law of the aircraft in service applies the various techniques of the verified control algorithm, such as dynamic inversion and eigenstructure assignment. The supersonic trainer employs the RSS(Relaxed Static Stability) concept in order to improve the aerodynamic performance in longitudinal axis and the longitudinal control laws employ the dynamic inversion with proportional-plus-integral control method. And, lateral-directional control laws employ the blended roll system of both beta-betadot feedback and simple roll rate feedback with proportional control method in order to guarantee aircraft stability. In this paper, the lateral-directional flight control law is designed by applying dynamic inversion control technique as a different method from the current supersonic trainer control technique, where the roll rate command system is designed at the lateral axis for the rapid response characteristics, and the sideslip command system is adopted at the directional axis for stability augmentation. The dynamic inversion of a simple 1st order model is applied. And this designed flight control law is confirmed to satisfy the requirement presented from the military specification. This study is expected to contribute to design the flight control law of KF-X(Korean Fighter eXperimental) which will proceed into the full-scale development in the near future.

• Proceedings of the Korean Institute of Navigation and Port Research Conference
• /
• v.1
• /
• pp.9-14
• /
• 2006

This paper develops a complete methodology for the mitigation of ionosphere spatial anomalies by GBAS systems fielded in the Conterminous U.S. (CONUS). It defines an ionosphere anomaly threat model based on validated observations of unusual ionosphere events in CONUS impacting GBAS sites in the form of a linear ‘wave front’ of constant slope and velocity. It then develops a simulation-based methodology for selecting the worst-case ionosphere wave front impact impacting two satellites simultaneously for a given GBAS site and satellite geometry, taking into account the mitigating effects of code-carrier divergence monitoring within the GBAS ground station. The resulting maximum ionosphere error in vertical position (MIEV) is calculated and compared to a unique vertical alert limit, or $VAL_{H2,I}$, that applies to the special situation of worst-case ionosphere gradients. If MIEV exceeds $VAL_{H2,I}$ for one or more otherwise-usable subset geometries (i.e., geometries for which the 'normal' vertical protection level, or $VPL_{H0}$, is less than the 'normal' VAL), the broadcast ${\sigma}_{pr_{-}gnd}$ and/or ${\sigma}_{vig}$ must be increased such that all such potentially-threatening geometries have VPL$_{H0}$ > VAL and thus become unavailable. In addition to surveying all aspects of the methods used to generate the required ${\sigma}_{pr_{-}gnd}$ and ${\sigma}_{vig}$ inflation factors for CONUS GBAS sites, related methods for deriving similar results for GBAS sites outside CONUS are suggested.

• Proceedings of the Korean Institute of Navigation and Port Research Conference
• /
• v.2
• /
• pp.235-240
• /
• 2006

Ionospheric scintillation induces a rapid change in the amplitude and phase of radio wave signals. This is due to irregularities of electron density in the F-region of the ionosphere. It reduces the accuracy of both pseudorange and carrier phase measurements in GPS/satellite based Augmentation system (SBAS) receivers, and can cause loss of lock on the satellite signal. Scintillation is not as strong at mid-latitude regions such that positioning is not affected as much. Severe effects of scintillation occur mainly in a band approximately 20 degrees on either side of the magnetic equator and sometimes in the polar and auroral regions. Most scintillation occurs for a few hours after sunset during the peak years of the solar cycle. This paper focuses on estimation of the effects of ionospheric scintillation on GPS and SBAS signals using a software receiver. Software receivers have the advantage of flexibility over conventional receivers in examining performance. PC based receivers are especially effective in studying errors such as multipath and ionospheric scintillation. This is because it is possible to analyze IF signal data stored in host PC by the various processing algorithms. A L1 C/A software GPS receiver was developed consisting of a RF front-end module and a signal processing program on the PC. The RF front-end module consists of a down converter and a general purpose device for acquiring data. The signal processing program written in MATLAB implements signal acquisition, tracking, and pseudorange measurements. The receiver achieves standalone positioning with accuracy between 5 and 10 meters in 2drms. Typical phase locked loop (PLL) designs of GPS/SBAS receivers enable them to handle moderate amounts of scintillation. So the effects of ionospheric scintillation was estimated on the performance of GPS L1 C/A and SBAS receivers in terms of degradation of PLL accuracy considering the effect of various noise sources such as thermal noise jitter, ionospheric phase jitter and dynamic stress error.

• KIPS Transactions on Software and Data Engineering
• /
• v.8 no.3
• /
• pp.123-128
• /
• 2019

This paper proposes the stock price prediction based on the artificial intelligence, where the model with recurrent convolution neural network (RCNN) layers is adopted. In the motivation of this prediction, long short-term memory model (LSTM)-based neural network can make the output of the time series prediction. On the other hand, the convolution neural network provides the data filtering, averaging, and augmentation. By combining the advantages mentioned above, the proposed technique predicts the estimated stock price of next day. In addition, in order to emphasize the recent time series, a custom weighted loss function is adopted. Moreover, stock data related to the stock price index are adopted to consider the market trends. In the experiments, the proposed stock price prediction reduces the test error by 3.19%, which is over other techniques by about 19%.

• Journal of the Korea Society for Simulation
• /
• v.31 no.2
• /
• pp.11-20
• /
• 2022

There is a critical need for AI assistance in guard operations of Army base perimeters, which is exacerbated by changes in the national defense and security environment such as force reduction. In addition, the possibility for human error inherent to perimeter guard operations attests to the need for an innovative revamp of current systems. The purpose of this study is to propose a real-time object detection AI tailored to military CCTV surveillance with three unique characteristics. First, training data suitable for situations in which relatively small objects must be recognized is used due to the characteristics of military CCTV. Second, we utilize a data augmentation algorithm suited for military context applied in the data preparation step. Third, a noise reduction algorithm is applied to account for military-specific situations, such as camouflaged targets and unfavorable weather conditions. The proposed system has been field-tested in a real-world setting, and its performance has been verified.