This study introduces a novel approach to address challenges in workpiece surface defect identification. It presents an enhanced Single Shot MultiBox Detector model, incorporating attention mechanisms and multi-feature fusion. The research methodology involves carefully curating a dataset from authentic on-site factory production, enabling the training of a model with robust real-world generalization. Leveraging the Single Shot MultiBox Detector model lead to improvements integrating channel and spatial attention mechanisms in the feature extraction network. Diverse feature extraction methods enhance the network’s focus on crucial information, improving its defect detection efficacy. The proposed model achieves a significant Mean Average Precision (mAP) improvement, reaching 99.98% precision, a substantial 3% advancement over existing methodologies. Notably, the proposed model exhibits a tendency for the values of the P-R curves in object detection for each category to approach 1, which allows a better balance between the requirements of real-time detection and precision. Within the threshold range of 0.2 to 1, the model maintains a stable level of precision, consistently remaining between 0.99 and 1. In addition, the average running speed is 2 fps lower compared to other models, and the reduction in detection speed after the model improvement is kept within 1%. The experimental results indicate that the model excels in pixel-level defect identification, which is crucial for precise defect localization. Empirical experiments validate the algorithm’s superior performance. This research represents a pivotal advancement in workpiece surface defect identification, combining technological innovation with practical efficacy.
1 September 2023 in Indonesian Journal of Electrical Engineering and Computer Science
Floods are a common natural hazard in Malaysia during the monsoon season. It affects millions of people each year that leads to severe deaths and infrastructure destruction. In recent time, flood warning system (FWS) has been a notable topic but it has not been extensively implemented in Malaysia. In this study, we developed a FWS that can interface with connected vehicles in order to provide alerts to drivers while also sending warnings to end users. This type of FWS enables vehicles to connect with one another within a particular radius to broadcast flood information via long range (LoRa) communication technology. When the water level rises over a certain point, the system sends a warning to drivers indicated through a mobile application. Drivers have the option to take alternative route, reducing the likelihood of damage when driving into or near a flooded area. The developed application demonstrates that the warning was able to be instantly displayed to the driver if there is a significant increase in water level. Experimental evidence shows that the driver was able to receive the water level alert and a visual interpretation of the immediate area affected by flood through the application.
1 November 2023 in Indonesian Journal of Electrical Engineering and Computer Science
The subject of fatigue monitoring is becoming more important in transportation and traffic management (including, for instance, the development of systems to detect and prevent driver drowsiness). People who work in offices are also susceptible to exhaustion, but there is currently no widely deployed system that is able to monitor this condition. In most cases, the driver’s eyelids will become heavy due to exhaustion after lengthy hours of driving or in the absence of mental concentration. Typically, when the driver’s concentration begins to fade, audio alert would be provided to force the drivers awake. In recent times, drowsiness is risky since it can result in an accident. Thus, a solution has been proposed to identify driver drowsiness by comparing several algorithms to find improved accuracy and execution time. Besides, this system will alert the driver with an audible warning in the event of drowsiness is detected.
Driver drowsiness and alcohol impairment are major causes of traffic accidents, making road safety a main concern. This study highlights the importance of addressing these issues through improved driver monitoring technologies. A prototype combining MQ-3 alcohol sensors, and facial detection was created, integrating with IoT via a Raspberry Pi to monitor and alert on drowsiness and alcohol levels. The developments use the NTHU-DDD dataset, which supports a supervised learning approach to develop a reliable drowsiness detection model. The study explored various machine learning algorithms such as Logistic Regression, Support Vector Machine (SVM), Random Forest (RF), K-nearest neighbors (KNN), Gradient Boosting Classifier, and Gaussian Naive Bayes, with Random Forest and Gradient Boosting emerging as top performers, particularly suited to complex non-linear data. The system effectively used supervised learning techniques to differentiate drowsy and non-drowsy images and exhibited consistent accuracy in detecting drowsiness, especially when the driver’s face was centered. However, accuracy decreased when faces were tilted, highlighting areas for refinement. Moreover, the environmental tests on the MQ-3 sensor demonstrated its sensitivity to alcohol presence, even distinguishing the intensity based on beverage type and concentration. The findings underscore the efficacy of using sensor-based technologies in real-world conditions and provide a foundation for optimizing the system's detection capabilities across various scenarios.
1 April 2023 in International Journal of Informatics and Communication Technology (IJ-ICT)
In recent times, the volume of traffic congestion has been rapidly growing on roads. These days the necessity of having safe transportation and journey is essential. Thus, vehicle communication could be a possible solution to enhance safe transit. Vehicular communications provide a wide range of applications with different characteristics, namely vehicle and vehicle (V2V) communications. Every year, traffic accidents kill many people worldwide, and many people have been injured. V2V communication enables vehicles to communicate with each other to provide safety and convenience to drivers. Therefore, this paper explores a direction to develop a conceptual approach to V2V communication with HC-12, nRF24L01, and XBee. The study aims to analyse and evaluate the communication range that may contribute to the future road transportation system.