Publication Search

The rapid proliferation of Internet of Things (IoT) devices across diverse industries has significantly increased the vulnerability of IoT edge networks to sophisticated cyber threats. Traditional intrusion detection systems (IDS), such as signature-based and anomaly-based approaches, are often insufficient in addressing the dynamic and evolving nature of these threats. This study proposes a hybrid intrusion detection system (IDS) framework that combines supervised machine learning (ML) techniques with deep reinforcement learning (DRL) to enhance detection performance in real-time, resource-constrained IoT environments. The proposed framework utilizes supervised learning for initial traffic classification and DRL for adaptive decision-making, enabling the system to continuously learn and optimize its detection policies based on new attack patterns. The hybrid approach significantly improves detection accuracy and reduces false positives when compared to conventional signature-based and single-model ML systems. In addition to improved detection capabilities, the framework's computational efficiency allows it to operate effectively within the constraints of IoT devices, ensuring that it is suitable for large-scale deployments. Benchmark evaluations using publicly available datasets, such as NSL-KDD, IoT-23, and BoT-IoT, show that the hybrid IDS framework outperforms traditional methods, providing a more robust and adaptive solution to cybersecurity challenges in IoT edge networks. The findings of this study suggest that combining machine learning with deep reinforcement learning offers a promising approach to secure IoT environments and address the limitations of existing IDS techniques. Future work will explore enhancing real-time adaptability, scalability, and the detection of zero-day attacks in evolving IoT ecosystems.

The city of Singkawang has experienced rapid population growth, leading to an increase in the number of students. On 17 September 2024, the Singkawang City Transportation Agency implemented a free revitalized Student Transport service to reduce the number of traffic accidents involving students. The aim of this study is to identify service and respondent characteristics, evaluate operational management, and assess performance and user satisfaction using the IPA and CSI methods, as well as Vehicle Operating Costs (VOC). The research employs a descriptive quantitative method by analyzing descriptive statistical data. Data were obtained from field observations, interviews, and surveys of 400 student respondents (146 users and 254 non-users). The results show that the student transport operates with two vehicles serving the North and East Singkawang routes. The load factor for outbound trips is 22%–32% and for return trips is 11%–12%, with travel times of 68–85 minutes, average operating speeds of 20–22 km/hour, and circulation times of 68–85 minutes. Based on the IPA analysis, the indicators in quadrant D require socialization regarding the functions and use of the interior of the student transport. The CSI result shows a score of 99.79% (very satisfied). The annual VOC amounts to IDR 292,905,814 (East Singkawang) and IDR 282,020,390 (North Singkawang). In conclusion, this service is satisfactory but still requires socialization to enhance its attractiveness and effectiveness.

Private vehicles are a frequently used mode of transportation because they are considered more practical. However, using private vehicles carries several risks, such as traffic accidents due to drivers losing focus on the road due to other activities, such as making calls on smartphones, drinking, or operating the radio. Approximately 90% of accidents are caused by human error. Convolutional Neural Network (CNN) is a type of neural network commonly used on image data. CNN is often used for image classification due to its high performance and accuracy. Therefore, this study aims to analyze the performance of CNN for the classification of distracted driving activities. The results show that the CNN model is able to effectively classify images of distracted driving activities, with an accuracy of approximately 99% across all datasets and across all input image size variations. Furthermore, the results of this study also show that differences in right-hand and left-hand drive datasets do not significantly affect model accuracy. Variations in input image size also do not significantly affect model accuracy, but do affect the training duration.

Traffic performance on urban road segments is strongly affected by vehicle volume and travel time, particularly during peak periods. This study analyzes the relationship between travel duration and the total number of vehicles passing along Otto Iskandar Road as an illustration of urban traffic conditions. Data were collected through field surveys, focusing on two main variables: average vehicle travel time and total traffic volume. Statistical analysis was performed using IBM SPSS Statistics, including normality testing and the Wilcoxon Signed Rank Test to identify potential differences between the observed variables. The results show a difference in average values between travel duration and vehicle volume; however, this difference is not statistically significant at the 95% confidence level (p = 0.180). These findings indicate that increases in traffic volume do not always lead to proportional increases in travel time, although they can still influence the stability and efficiency of traffic flow. The results are consistent with previous studies, such as Halim (2021), who reported that U-turn movements affect speed and traffic performance, and Handayani et al. (2024), who found that parking activities and vehicle maneuvers reduce road capacity. Other studies also highlight the impact of side friction and traffic flow variations on speed and saturation levels. Overall, this study emphasizes the importance of managing vehicle flow and monitoring travel time in urban transportation planning and traffic management.

Increasing traffic activity at unsignalized intersections often results in congestion and delays. The intersection of Jl. Imam Bonjol and Jl. Daya Nasional experiences these conditions, making performance evaluation necessary. This research tends to scrutinize the intersection performance under existing conditions and five-year projected conditions and to identify appropriate improvement measures. The analysis was carried out using the 2023 Indonesian Road Capacity Guidelines (PKJI 2023) and PTV VISSIM microsimulation software. Data on intersection geometry, vehicle speed, and traffic volume were collected through CCTV observations on weekdays and weekends. The results show that under existing conditions, the degree of saturation reached 1.170 with an average delay of 33.75 seconds (LOS D). In the five-year projection, the degree of saturation increased to 1.239 with an average delay of 53.69 seconds (LOS E). These findings indicate a decline in intersection performance, highlighting the need for alternative traffic management measures to improve operational performance and service levels.    

The rapid development of the Internet of Vehicles (IoV) has introduced new security challenges, particularly in protecting Controller Area Network (CAN Bus) communications from cyberattacks such as Denial of Service (DoS) and spoofing attacks. This study proposes the implementation of the Extreme Gradient Boosting (XGBoost) algorithm combined with Information Gain feature selection to improve intrusion detection performance in IoV environments. The CICIoV2024 dataset, which represents both benign and malicious traffic, is used as the primary data source. The research process includes data integration, preprocessing, feature selection, data splitting, and model training using a 5-fold cross-validation approach. Experimental results demonstrate that the proposed model achieves outstanding performance, with accuracy, precision, recall, and F1-score exceeding 99.99%, and an Area Under Curve (AUC) value approaching 1.00. Furthermore, Information Gain successfully identifies the most influential CAN payload features, enhancing model efficiency without sacrificing accuracy. These findings confirm that the combination of Information Gain and XGBoost is highly effective for developing a fast, accurate, and efficient intrusion detection system in IoV networks.

The massive adoption of Internet of Things (IoT) devices is expanding the cyberattacks surface, particularly by the Mirai botnet, which exploits the dynamic characteristics of data traffic. This research proposes a Mirai detection approach based on a Recurrent Neural Network (RNN) optimized using Bayesian Optimization to improve prediction accuracy on sequential data. Unlike previous studies, this research utilizes the latest CIC IoT-DIAD 2024 dataset and applies probabilistic optimization to the hyperparameter space, including RNN units, dropout, and learning rate. The experiment was conducted on 201,021 valid data points, with dimensionality reduction using PCA as the optimal point to represent essential features without redundancy. The results show a significant increase in accuracy from 97.95% to 99.69%, accompanied by an 84% decrease in False Negatives, an 86% decrease in False Positives, and an AUC value of 0.9999. These findings confirm that integrating RNN and Bayesian Optimization not only improves numerical performance but also strengthens the reliability of the intrusion detection system for modern IoT ecosystems with controlled computational loads.

Traffic congestion is a persistent challenge in urban areas in Indonesia, where increasing vehicle density creates the need for intelligent traffic monitoring systems. This study aims to develop a real-time vehicle parking system using the YOLOv8 object detection model to provide efficient traffic analysis from live CCTV broadcasts and recorded videos. This study uses a quantitative experimental approach with the implementation of the YOLOv8m model using the Ultralytics library in Python, tested on data collected from CCTV cameras A TCS Dishub Medan and additional footage from mobile devices. Vehicles are detected and counted in two directions up (Up) and down (Down) using virtual detection lines on the video frame. The system performance is evaluated by automatic detection counting with manually recorded ground truth data. The results show that on live CCTV broadcasts, the YOLOv8m model achieves an average precision of 98.96%, a recall of 96.59%, and an F1 score of 97.74% for upstream traffic, while for downstream traffic it achieves 100% precision, 95.64% recall, and an F1 score of 97.730/0. On the other hand, on high-quality recorded videos, all performance metrics achieve 100%, indicating perfect detection accuracy. These findings confirm the effectiveness of YOLOv8 in real-time traffic monitoring, but also indicate that video quality and stream stability affect detection performance. In conclusion, the developed system shows strong potential to support smart city traffic management solutions. Future research should focus on performance optimization under low-resolution live streaming conditions to improve accuracy in practical applications.  

This qualitative study examines the Public Service Performance Gap at the Ponorogo Regency Transportation Agency (Dishub) in managing high-risk traffic assets, which stems from the failure to synergise Normative Governance with operational Public Asset Management (MAP). Utilising Edwards III's Policy Implementation Model and GG/NPS principles, the core finding indicates that synergy failure is mediated by a Reactive Bureaucratic Disposition. While Dishub adheres procedurally, asset maintenance is largely reactive—performed only after damage or public complaint—not preventive. This non-responsive attitude limits accountability to reporting outputs disconnected from physical service outcomes, leading to low service quality. The proposed substantive solution is to activate Community Involvement (NPS) as a key moderator, which is currently weak, by integrating Functional Participation into the agile MAP cycle. The research recommends an e-governance system with KPIs, where transparently integrated citizen reports automatically trigger work orders, creating external public accountability pressure that forces the reactive bureaucracy to act proactively.

The growth in the number of vehicles in Surabaya has led to increased traffic congestion, particularly on the Perak Timur–Perak Barat road section, which serves as a major distribution route to the port and industrial areas. Problems arise at U-Turn points, where vehicle accumulation hampers traffic flow. This study analyzes traffic volume, travel time, queue length, and queue ratio at two U-Turn points on this section. The method used was a field survey with direct observation of vehicle volume every 15 minutes, vehicle classification, and queue length. Traffic volume was calculated using the Passenger Car Equivalent (PCE) factor based on the 1997 MKJI standards. The study results show that at Point 1, east–west direction, the highest queue ratio occurred on Tuesday from 11:00 to 12:00 (p = 4.84), while the lowest was on Sunday from 06:00 to 07:00 (p = 0.21). At Point 2, west–east direction, the highest queue ratio occurred on Tuesday from 08:00 to 09:00 (p = 6.18), and the lowest from 06:00 to 07:00 (p = 0.41). These findings indicate that during peak hours, traffic congestion increases (p > 1), causing long queues, particularly in the west–east direction in the morning. The performance of the U-Turn on the Perak Timur–Perak Barat road section needs improvement, with recommendations such as temporarily closing U-Turns during high volumes, providing alternative U-Turn lanes, and adding signs to minimize the potential for vehicle conflicts.

Sea transportation plays a crucial role in supporting national development, particularly in transporting goods and passengers. Tanjung Emas Port in Semarang, as a major port with high traffic, faces significant challenges in ensuring navigational safety. This study aims to analyze the partial effects of three key operational factors—Communication System (X1), Safety Facilities (X2), and Crew Competence (X3)—on Navigational Safety (Y). A quantitative approach was used with a sample of 100 ship crew members, selected using incidental sampling. Data were analyzed through descriptive statistics and Multiple Linear Regression (MLR) using SPSS. The results indicate that all three independent variables have a positive and significant effect on navigational safety. The regression equation obtained, Y = -1.398 + 0.218X1 + 0.278X2 + 0.302X3 + μ, shows that Crew Competence (X3) is the most dominant variable in predicting navigational safety outcomes. Collectively, these variables explain 72.4% of the variance in navigational safety. These findings emphasize that human competence and adherence to procedural discipline are critical leverage points in improving safety performance at ports.

This study aims to analyse the design and optimization of digital marketing to increase the brand awareness of the IdRink MSME through the integration of a website and Instagram. The research employed a Research and Development (R&D) method using the Four-D model (Define, Design, Develop, Disseminate), combined with prototyping techniques and evaluation through blackbox testing. Data were collected through interviews, literature review, observation, Google Analytics, Instagram Insight, and A/B testing questionnaires administered to 100 respondents. The results show that the development of a WordPress-based website and a structured Instagram content plan significantly improved IdRink’s digital performance. Within one month, the audience growth rate increased by 256%, average post reach reached 216%, and website traffic recorded 298 users. Brand awareness measurement indicated brand recognition of 74% and brand recall of 79%, higher than the comparison brand in the local beverage category. These findings indicate that integrating a website and Instagram as a digital marketing strategy is effective in moving IdRink from the unaware-of-brand stage to recognition and recall, expanding market reach, strengthening customer engagement, and building a more professional brand image among young urban consumers. The study recommends maintaining content consistency and utilising digital analytics to support future marketing decision-making.

In the modern computing era, servers face significant challenges in data storage due to hardware failures, cyber attacks, or human errors. The problem highlighted focuses on the impact of file systems on three critical aspects: data integrity (accuracy and consistency of data without corruption), data recovery (the ability to restore data after a failure), and failure resilience (fault tolerance, such as redundancy and journaling to prevent downtime). The main issue is that traditional file systems like FAT32 or NTFS are often susceptible to fragmentation, metadata loss, or long recovery times, which can lead to data loss of up to 20-30% on enterprise servers, especially in high-traffic environments like cloud computing.A simple problem-solving process is conducted through a straightforward comparative analysis approach: (1) A literature review of popular file systems (ext4, ZFS, Btrfs); (2) Failure simulations using tools like fsck and stress testing on virtual servers (e.g., via KVM or Docker); and (3) Measuring performance metrics with benchmarking tools like Bonnie++ for I/O throughput, recovery time, and error rates. This process is designed to be simple, requiring only a virtual lab setup without expensive hardware, and is analyzed quantitatively with descriptive statistics.The solution to the problem indicates that advanced file systems like ZFS or Btrfs provide significant improvements: data integrity is up to 95% more secure through automatic checksums, data recovery is achieved in minutes through snapshots and RAID integration, and failure resilience is higher with copy-on-write features. The main recommendation is to migrate to journaling-based file systems for servers, combined with automated backups, which can reduce the risk of downtime by up to 50%. This research provides practical guidance for system administrators to enhance server reliability without excessive additional costs.

Container loading and unloading procedures are a series of essential operational activities carried out to systematically, safely, and efficiently move containers from ships to the stacking yard or vice versa. This study aims to examine these procedures, analyze the effectiveness of their implementation, and identify the obstacles and corrective efforts encountered in the field. The methods used include field observation (field research) and literature review (library research). The results indicate that the procedural stages involve document preparation (Bill of Lading and Manifest), ship berthing, the utilization of various loading and unloading equipment (such as Gantry Crane, RTG, and Headtruck), and container arrangement within the Container Yard (CY). The process involves three main work stages: Stevedoring, Cargodoring, and Receiving/Delivery, which require close coordination among port operators, foremen, and Stevedoring Workers (TKBM). Implementation can be hindered by crane breakdowns, headtruck limitations, traffic congestion, and slow stevedoring performance. Effective solutions suggested include routine equipment maintenance, increasing HT fleet capacity, enhancing human resource competency through training, and coordinating traffic routes with relevant authorities. Optimal procedure execution is crucial for ensuring smooth logistics flow, time efficiency, and occupational safety.

This study investigates the implementation strategy of work discipline as a means to improve the productivity of daily contract employees in the Traffic Division of the Semarang City Transportation Agency. Work discipline is a critical component of human resource management that influences performance effectiveness and organizational efficiency. The research applies a qualitative descriptive approach using interviews, observation, and document analysis to examine discipline practices, attendance data, and managerial supervision. The results show that effective disciplinary implementation—through structured supervision, reward and punishment systems, continuous coaching, and welfare enhancement—significantly improves employee productivity. Leadership quality, motivation, and consistent policy enforcement emerge as primary determinants of disciplined behavior. The study concludes that establishing a fair and transparent disciplinary system can foster accountability, punctuality, and professional service delivery among public employees.

PT Intelix, a technology and information company with numerous collaborative projects, faces challenges in effective project management and monitoring. Despite utilizing project management applications like Jira, the generation of final reports often remains a manual process using Microsoft Excel, which is time-consuming and prone to errors. To address these issues, PT Intelix developed an internal Project Management application, yet it required further enhancements in its dashboard and reporting features for more detailed project monitoring optimization. This research aims to develop the Dashboard menu by adding a priority project table and several sub-menus in the Report menu to accelerate and simplify the preparation of final project reports. The development methodology employed is Agile Kanban, encompassing planning, system and interface design, incremental development and implementation, and testing. System functionality testing was conducted using Black Box Testing and User Acceptance Test (UAT) involving the project owner and end-users from PT Intelix. Test results showed all main functions running smoothly and stably, with all black box scenarios passing. UAT testing on six menus received a "very good" response from the project owner and users. The SUS evaluation of three users scored a 5 on the scale, with several scoring a 4, indicating a good system. Performance testing with 10 simultaneous users also demonstrated stable traffic and no errors.. The development of High Priority Project, Milestone Project, and Implementation Project features on the Dashboard successfully presented more detailed project information, while the addition of sub-menus in the Report accelerated report generation. Thus, this development successfully optimized integrated project monitoring and reporting at PT Intelix, reducing manual effort and enhancing data accuracy for decision-making.

The increase in traffic volume in urban areas often triggers various problems, especially at Road intersection points that have a high level of interaction between vehicles. One of the locations that experience these problems is the intersection with the island of traffic without signal on Jl. Tebu-Jl. Tabrani Ahmad, City Of Pontianak. This study aims to analyze the performance of the intersection in the existing condition and project its performance in the next five years (2030), as well as formulate alternative treatment that can improve the smoothness of traffic. The analysis method refers to Indonesian road capacity guidelines (PKJI) 2023 and modeling using software PTV VISSIM. The data collected include intersection geometry, traffic volume, and vehicle speed. Traffic volumes were obtained through CCTV recordings for three consecutive days, then processed and presented in the form of tables and graphs for easy interpretation. The results of the existing condition analysis showed that the degree of saturation (DS) reached 0.407 with an average delay of 9.539 seconds (Level of Service/LOS B) based on PKJI, while VISSIM simulation resulted in a delay of 12.54 seconds (LOS B). The five-year projection (2030) shows an increase in DS to 0.878 with a delay of 15.177 seconds (LOS C) from PKJI, and 23.35 seconds (LOS C) from VISSIM, which indicates a decrease in Junction performance as traffic grows. Two handling alternatives were analyzed, namely the construction of roundabouts and the implementation of traffic flow management. VISSIM simulation shows that the roundabout can reduce the delay to 14.31 seconds with a queue length of 20.34 m (LOS B), while the flow management produces a delay of 11.81 seconds with a queue of 13.14 m (LOS B). This result confirms that both alternatives are able to improve the performance of the intersection compared to the projected condition without handling, so that it can be a technical recommendation for urban traffic planning in Pontianak.

Improvement amount Distributed Denial of Service (DDoS) attacks in cloud infrastructure and edge computing demands solution adaptive, distributed, and efficient detection in a way computing. Research This propose an optimized Federated Learning (FL) based DDoS detection model using Centroid Opposition-Based Bacterial Colony Optimization (COBCO) to training the Elman Neural Network (ENN). The proposed architecture consists of of two components Main: on the edge node side, a hybrid Convolutional Neural Network–Gated Recurrent Unit (CNN–GRU) model is used to extraction feature local from traffic data network, while on the server side, model parameters from each node are collected and used for training an optimized ENN with COBCO. Approach This aim increase accuracy detection at a time maintain efficiency local data communication and privacy. In progress experimental, model tested use three benchmark datasets: NSL-KDD, CICIDS2017, and CICDDoS2019. The preprocessing process includes feature encoding categorical, normalization numeric, class balancing using SMOTE, as well as validation cross (k-fold). Initial results show that combination of FL, CNN–GRU, and COBCO–ENN produces improvement significant in accuracy and time convergence compared to approach conventional such as PSO, GA, and non- federative models. In addition, the proposed model capable maintain performance detection tall although executed in edge environment with limitations source Power.  Study This give contribution important in development system scalable, privacy-preserving, and adaptive intelligent DDoS detection to dynamics Then cross modern network. Integration of FL and COBCO in ENN training shows potential big for used in implementation real in cloud-edge infrastructure. In addition, the proposed model demonstrates strong scalability and adaptability, making it highly suitable for dynamic and evolving network environments.

The digital transformation in education has encouraged the adoption of computer-based tests (CBT) using video content, which demands stable and efficient network performance. This study aims to evaluate the performance of two queue management algorithms, namely Random Early Detection (RED) and Per Connection Queue (PCQ), in maintaining the quality of service (QoS) of school networks during online video-based examinations. A case study approach was applied using a real network topology in a school environment, and QoS parameters such as throughput, delay, packet loss, and jitter were measured. The implementation was conducted using a MikroTik RB450Gx4 router configured with simple queue settings for each algorithm. The results show that PCQ provides more consistent performance under high user loads, achieving an average throughput of 56,482 bps and lower delay compared to RED. Conversely, RED performs better in scenarios with a small number of users. The study recommends using PCQ for networks with dynamic and dense user environments, while RED is more suitable for low-traffic conditions where latency stability is crucial. These findings offer practical guidance for managing bandwidth and improving the quality of CBT delivery in educational settings.

This research focuses on the development and evaluation of an Adaptive Edge-AI framework designed to optimize real-time data processing and decision-making in resource-constrained environments, specifically within smart city infrastructures. The primary problem addressed is the challenge of minimizing latency, reducing energy consumption, and ensuring the reliability of Cyber-Physical Systems (CPS) when using Internet of Things (IoT) devices. The objective of the study is to assess the effectiveness of this framework in real-world smart city applications such as traffic monitoring, environmental sensing, and smart utilities management. The proposed method integrates lightweight AI models, edge computing, and adaptive resource management techniques, including Federated Learning and Neural Architecture Search, to ensure optimal performance while addressing hardware constraints. The main findings reveal that the framework significantly improves real-time inference speed, reduces energy consumption of IoT devices, and enhances CPS reliability by minimizing communication delays and ensuring continuous system operation despite network disruptions. The application of this framework to smart transportation and urban utilities further demonstrates its potential to optimize city management processes. The study concludes that the Adaptive Edge-AI framework offers a promising solution for smart cities, enhancing operational efficiency, sustainability, and resilience. It is recommended for integration into smart city infrastructures to enable better resource management and decision-making in real-time applications.