Byungmun Kang

This study introduces Bot Fit, an innovative hip-joint exoskeleton designed for the objective assessment of lower limb muscular strength and exercise performance. A total of 25 participants underwent baseline strength assessments using conventional fitness equipment before performing resistance exercises under two controlled conditions: a 10-second test without speed restrictions and a 30-second test at a constant speed. Bot Fit recorded motor signals and performance metrics, capturing key indicators such as repetitions and movement patterns. To analyze the relationship between muscle parameters and exercise performance, we employed correlation analysis, hierarchical cluster analysis (HCA), and regression analysis. The results demonstrated strong associations between lower limb strength and key exercises, including squats, knee-ups, and reverse lunges, particularly in short-duration and constant-speed conditions. HCA successfully clustered participants based on three exercise-related metrics with an accuracy of 0.89, while the regression model achieved a correlation coefficient of 0.86, indicating high statistical power. These findings suggest that Bot Fit provides a reliable, data-driven method for muscular strength assessment, offering valuable insights for optimizing exercise programs, preventing injuries, and enhancing physical performance. The integration of wearable robotic technology into health and fitness monitoring enables personalized exercise interventions and supports evidence-based decision-making in sports science, rehabilitation, and performance enhancement.

Anomaly detection has been a challenging subject in many industrial fields. In industrial machinery such as hydraulic excavators, sensor data distributions are inherently multimodal because different operating conditions produce distinct sensor signatures, and conventional algorithms struggle to establish clear normal–abnormal boundaries when these conditions are mixed. We propose an action-state decomposition framework that partitions multimodal sensor data into homogeneous subsets based on discretized control inputs, thereby reducing the ambiguity of normal–abnormal boundaries by learning state-conditional distributions. The approach comprises a reactive method that evaluates each sample within its action state, and a history-based method that incorporates temporal context from previous action states. This decomposition is algorithm-agnostic and can improve detection performance across diverse anomaly detection algorithms. The framework is further extended to Bayesian fault diagnosis that identifies the root cause of failures using action-state-conditional detection probabilities. Experiments on simulated excavator data and two real-world benchmark datasets (UCI Hydraulic Systems and SKAB) demonstrate the generalizability of the proposed mode decomposition and provide insights into factors that may influence its effectiveness. The history-based method achieves a mean AUC of 0.89 across sensor fault types, outperforming all baseline algorithms, and the Bayesian fault diagnosis achieves 86.7% accuracy in identifying the root cause among six action fault types. For the proposed GMM-based methods, the decomposition also substantially reduces per-sample inference time by approximately 10× (from 8.68 μs to 0.75 μs), enabling real-time deployment in industrial settings.

Electrocardiogram waveform delineation is a fundamental task for quantitative cardiac analysis, yet accurate and consistent estimation of waveform boundaries remains challenging due to heart rate variability, inter-subject morphological differences, and nonlinear temporal distortions across cardiac cycles. Conventional rule-based methods and pointwise Dynamic Time Warping approaches are sensitive to amplitude variations and baseline fluctuations, while deep learning–based models require large annotated datasets and often suffer from limited interpretability and generalization. In this study, we propose a morphology-oriented ECG waveform alignment framework based on Pearson correlation–based Dynamic Time Warping (PCDTW). By integrating window-level matching with a correlation-driven cost function, the proposed method explicitly emphasizes local morphological similarity rather than absolute amplitude differences. Each ECG record is aligned using a subject-specific reference cycle constructed from normalized RR intervals, enabling stable correspondence of waveform boundaries without any training process. The proposed method was evaluated on two publicly available databases, the QT Database (QTDB) and the Lobachevsky University Electrocardiography Database (LUDB). Experimental results show that PCDTW significantly reduces QT and QTcB estimation errors compared with conventional DTW variants, demonstrating improved temporal consistency and lower bias across cardiac cycles. In particular, the mean QTcB error was reduced to 28.14 ms, compared with 124.54 ms obtained using conventional DTW. In addition, on LUDB, the overall mean delineation error for the P wave, QRS complex, and T wave boundaries was 10.68 ms, showing comparable or superior performance to state-of-the-art deep learning–based methods despite requiring no external training data. These findings indicate that morphology-aware, correlation-based temporal alignment provides a robust and interpretable alternative for ECG waveform boundary detection under realistic physiological variability.

Moving object detection and tracking are technologies applied to wide research fields including traffic monitoring and recognition of workers in surrounding heavy equipment environments. However, the conventional moving object detection methods have faced many problems such as much computing time, image noises, and disappearance of targets due to obstacles. In this paper, we introduce a new moving object detection and tracking algorithm based on the sparse optical flow for reducing computing time, removing noises and estimating the target efficiently. The developed algorithm maintains a variety of corner features with refreshed corner features, and the moving window detector is proposed to determine the feature points for tracking, based on the location history of the points. The performance of detecting moving objects is greatly improved through the moving window detector and the continuous target estimation. The memory-based estimator provides the capability to recall the location of corner features for a period of time, and it has an effect of tracking targets obscured by obstacles. The suggested approach was applied to real environments including various illumination (indoor and outdoor) conditions, a number of moving objects and obstacles, and the performance was evaluated on an embedded board (Raspberry pi4). The experimental results show that the proposed method maintains a high FPS (frame per seconds) and improves the accuracy performance, compared with the conventional optical flow methods and vision approaches such as Haar-like and Hog methods.