Magnetic compass based on MEMS is a device that can display the course of ships and the heading of aircraft, and it is widely used because of its fast measurement and low cost.

Due to the characteristics of Earth’s magnetic field and the environmental effects on magnetic compasses, determination of heading based on Earth’s magnetic field measurement still exposes several problems to overcome. Thus, lots of efforts have been made to improve the accuracy of digital magnetic compasses in different external conditions, which resulted in the development of many calibration methods.

Preceding research papers described that the locus from the sensing axes of magnetic compass is an ellipse or ellipsoid due to the influence of magnetic interference on the magnetic compass.

Ri Yong Rim, a researcher at the Faculty of Naval Architecture and Ocean Engineering, proposed a method for compensating the accidental error and improving the accuracy of MEMS magnetic compass by using robust ellipse fitting.

In order to evaluate the effectiveness of the objective function, he performed ellipse approximations by the measured values under artificial random noise. As a result, he identified the trend and shortcomings of the new objective function, and further improved the objective function to overcome the shortcomings.

Then, he obtained compensation coefficients and compensation errors by the algebraic ellipse approximation and by the improved ellipse approximation, calculated the heading, and compared it with the actual heading. The result demonstrated that the improved ellipse approximation model had good noise suppression performance and adaptability.

You can find more information in his paper “Research on Improving Heading Accuracy of MEMS Magnetic Compass Based on Robust Ellipse and Spheroid Fitting” in “Proceedings of KUTIC-2025”.

Today, models of cooperative games with transferable utility (TU-games for short) have been widely used in many fields including economics. One of the most important problems in cooperative TU-games is to determine rational imputation, and computing such imputation within a limited time is of practical importance. So far, cooperative TU-games have used the Shapley value and the nucleolus to determine imputation, but the problem of computing the nucleolus of classical TU-games is NP-hard. In addition, the Shapley value computation in classical cooperative TU-games requires exponential time.

GWO (Grey Wolf Optimizer) is an algorithm to solve optimization problems by simulating the foraging process of grey wolves.

O Un Suk, a lecturer at the Faculty of Applied Mathematics, proposed an algorithm for computing rational imputation (an element of the core) in TU-games with nonempty core within a limited time.

The calculation experimental results demonstrated that the proposed algorithm is much better than the calculation of the Shapley value if the number of players exceeds four in classical TU-games.

If more details are needed, please refer to her paper “Algorithm for Computing the Rational Imputation using Grey Wolf Optimizer in Cooperative TU-Games” in “Proceedings of KUTIC-2025”.

The magnetometer is one of the most suitable attitude sensors for almost all low earth orbit (LEO) satellites.

There have been lots of efforts to develop attitude estimation algorithms using only a magnetometer and a spacecraft attitude estimation algorithm utilizing only magnetometer readings was developed. Although the algorithm can provide high accuracy of satellite attitude estimation, it is widely known that the unscented Kalman filter (UKF) faces heavy computational load.

Kim Hak Min, a researcher at the Faculty of Aerospace Engineering, proposed an attitude and angular rate estimation algorithm based on quaternion estimation (QUEST) and unscented Kalman filter (UKF).

The proposed algorithm uses only three-axis magnetometer as an attitude sensor and works in two stages. In the first stage, the QUEST uses magnetometer measurement vector and its derivative to estimate a coarse attitude. In the second stage, these estimated values are used as quaternion measurement data for UKF processing. The state vector for UKF consists of attitude parameters and angular rate vector. As a result, it is possible to estimate satellite attitude and get angular rate with high accuracy and small computation load.

The simulation result showed that the attitude of a satellite can be estimated with higher accuracy than 0.2° with small computational load.

You can find the details in his paper “Research on LEO Satellite Attitude Estimation Based on Quaternion and Unscented Kalman Filter” in “Proceedings of KUTIC-2025”.