KJCKorea
Journal Central

This paper proposes the design and implementation of an intelligent automatic speed bump system incorporating Internet of Things (IoT) technology. Conventional fixed speed bumps have limitations in terms of road safety and vehicle protection, including noncompliance with installation standards, unnecessary impact on low-speed vehicles, and delays in the movement of emergency vehicles. To address these issues, this study implemented an automatic speed bump system that integrates a CSI camera, load cells, and a microwave motion sensor (TRM-121A) based on Jetson Nano and Arduino Uno. The proposed system determines whether a vehicle is speeding by simultaneously measuring section speed using load cells and instantaneous speed using the motion sensor. It also identifies general vehicles and emergency public service vehicles, such as ambulances and police cars, using a lightweight YOLO algorithm. When an emergency vehicle is detected, the bump remains flat to support rapid passage. When a general vehicle exceeds the speed limit, the actuator is controlled to activate the speed bump function. In addition, a web-based monitoring system using Flask and AWS was developed to enable real-time road condition monitoring, speed threshold adjustment, and manual control of the speed bump.

The proliferation of the Internet of Things (IoT) environment has led to the generation of diverse sensor-based data, and research on intelligent classification and prediction models utilizing this data is actively underway. However, in real IoT application environments, acquiring large-scale data is often difficult due to factors such as data collection costs, security concerns, and equipment limitations. In these small-scale data environments, it is challenging to meet the vast training data requirements of deep learning-based models, and overfitting issues can become severe. The objective of this study is to experimentally analyze the generalization performance and stability of Random Forest in small-scale IoT data environments. To achieve this, comparative experiments are conducted using the Small IoT Dataset provided by Kaggle, evaluating Random Forest, Support Vector Machine (SVM), and neural network models. Specifically, by applying cross-validation-based evaluation instead of a single training-validation split, we quantitatively compare the generalization capabilities of each model by analyzing not only their average performance but also their performance variance. This analysis is expected to contribute to the application of machine learning in rapidly changing small-scale IoT environments.

As the importance of visual presentation and consumer experience in sports brand stores continues to increase, there is a growing need to examine the effects of Visual Merchandising (VMD) components on consumers’ emotional responses and purchase intentions. The purpose of this study is to compare and analyze the effects of VMD components in Chinese and international sports brand stores on consumers’ emotional responses and purchase intentions. To achieve this objective, LI-NING, representing a Chinese brand, and Nike, representing an international brand, were selected as the research subjects. An online survey was conducted among female college students in their twenties residing in Changchun, China, from July 1 to July 10, 2025, and a total of 350 valid questionnaires were used for the final analysis. The collected data were analyzed using SPSS 26.0. The results indicated that, in both Chinese and international sports brand stores, the VMD components namely spatial elements, presentation elements, and display elements had significant positive effects on consumers’ emotional responses and purchase intentions. In addition, consumers’ emotional responses, including pleasure, arousal, and dominance, were found to have significant positive effects on purchase intention. This study has both academic and practical implications in that it compares and analyzes the effects of VMD components in Chinese and international sports brand stores on consumers’ emotional responses and purchase intentions, thereby providing fundamental insights for establishing effective VMD strategies in sports brand stores.