Gregorius Airlangga

Mr. Gregorius Airlangga


Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
Verified
Journal Article
Open access
PDF file
GAGregorius Airlangga
Gregorius Airlangga
ALAlan Liu
Alan Liu
Urban happiness prediction presents a complex challenge, due to the nonlinear and multifaceted relationships among socio-economic, environmental, and infrastructural factors. This study introduces an advanced hybrid model combining a gradient boosting machine (GBM) and neural network (NN) to address these complexities. Unlike traditional approaches, this hybrid leverages a GBM to handle structured data features and an NN to extract deeper nonlinear relationships. The model was evaluated against various baseline machine learning and deep learning models, including a random forest, CNN, LSTM, CatBoost, and TabNet, using metrics such as RMSE, MAE, R2, and MAPE. The GBM + NN hybrid achieved superior performance, with the lowest RMSE of 0.3332, an R2 of 0.9673, and an MAPE of 7.0082%. The model also revealed significant insights into urban indicators, such as a 10% improvement in air quality correlating to a 5% increase in happiness. These findings underscore the potential of hybrid models in urban analytics, offering both predictive accuracy and actionable insights for urban planners.

25 April 2023 in Drones
Verified
Journal Article
Open access
PDF file
GAGregorius Airlangga
Gregorius Airlangga
ALAlan Liu
Alan Liu
An unmanned aerial vehicle (UAV) is an autonomous flying robot that has attracted the interest of several communities because of its capacity to increase the safety and productivity of labor. In terms of software engineering, UAV system development is extremely difficult because the focus is not only on functional requirement fulfillment, but also on nonfunctional requirements such as security and safety, which play a crucial role in mission success. Consequently, architecture robustness is very important, and one of the most common architectures developed is based on a centralized pattern in which all UAVs are controlled from a central location. Even though this is a very important problem, many developers must expend a great deal of effort to adapt and improve security. This is because there are few practical perspectives of security development in the context of UAV system development; therefore, the study of attack and defense patterns in centralized architecture is required to fill this knowledge gap. This paper concentrates on enhancing the security aspect of UAV system development by examining attack and defense patterns in centralized architectures. We contribute to the field by identifying 26 attack variations, presenting corresponding countermeasures from a software analyst’s standpoint, and supplying a node.js code template for developers to strengthen their systems’ security. Our comprehensive analysis evaluates the proposed defense strategies in terms of time and space complexity, ensuring their effectiveness. By providing a focused and in-depth perspective on security patterns, our research offers crucial guidance for communities and developers working on UAV-based systems, facilitating the development of more secure and robust solutions.

29 February 2024 in MALCOM: Indonesian Journal of Machine Learning and Computer Science
Verified
Journal Article
Open access
PDF file
GAGregorius Airlangga
Gregorius Airlangga
This research investigates the performance of various machine learning models in developing an Intrusion Detection System (IDS) for the complex and evolving security landscape of Internet of Things (IoT) networks. Employing the RT-IoT2022 dataset, which captures a diverse array of IoT devices and attack methodologies, we meticulously evaluated four prominent models: Gradient Boosting, Random Forest, Logistic Regression, and Multi-Layer Perceptron (MLP). Our results indicate that both Gradient Boosting and Random Forest achieved perfect scores with an accuracy, precision, recall, and F1 score of 1.00, suggesting their superior ability to classify and predict security incidents within the dataset. Logistic Regression demonstrated commendable consistency with scores of 0.96 across all metrics, proposing a balance between model complexity and performance. The MLP model closely followed, with an accuracy, precision, recall, and F1 score of 0.99, highlighting its potential in capturing complex, nonlinear data relationships. These findings underscore the critical role of machine learning in fortifying IoT networks against cyber threats and the need for continuous model evaluation against real-world data. The study provides a pathway for future research to refine these IDS models for operational efficiency and sustainability in the dynamic IoT security domain.

4 April 2022 in IEEE Access
Verified
Journal Article
Open access
GAGregorius Airlangga
Gregorius Airlangga
ALAlan Liu
Alan Liu
A motion strategy plays an important role in supporting autonomous Unmanned Aerial Vehicle (UAV) movement. Many studies have been conducted to improve the motion frameworks in terms of its robustness, safety and performance. Most of them worked on the prior known maps scenario where the area information was collected by Global Positioning System (GPS) and satellite cameras. Even though the scheme can provide high quality map, the computation of motion planning remains dependent on the communication signal. In the rural areas such as forests and mountains, where communication signal does not perform well, unclear and noisy terrain maps can be generated and lead to mission failure. Therefore, it is significant that an alternative framework to enhance autonomous UAV motion performance in these certain conditions should be developed. Our work focuses on developing a high performance path planner for autonomous UAV motion when communication signal does not work well in rural areas. The search mission problem in forest terrain has been implemented in 3D simulation as an evaluation. By conducting a simulation process repeatedly with different test cases for positions, time constraints, flight speed (3-11 m/s) and flight range, our path planning framework can achieve completeness between 90-100% and better performance compared to others.

20 December 2024 in Designs
Verified
Journal Article
Open access
PDF file
GAGregorius Airlangga
Gregorius Airlangga
RSRonald Sukwadi
Ronald Sukwadi
WBWidodo Widjaja Basuki
Widodo Widjaja Basuki
LSLai Ferry Sugianto
Lai Ferry Sugianto
Oskar Ika Adi NugrohoOskar Ika Adi Nugroho
Oskar Ika Adi Nugroho
RRRadyan Rahmananta
Radyan Rahmananta
This study evaluates and compares the computational performance and practical applicability of advanced path planning algorithms for Unmanned Aerial Vehicles (UAVs) in dynamic and obstacle-rich environments. The Adaptive Multi-Objective Path Planning (AMOPP) framework is highlighted for its ability to balance multiple objectives, including path length, smoothness, collision avoidance, and real-time responsiveness. Through experimental analysis, AMOPP demonstrates superior performance, with a 15% reduction in path length compared to A*, achieving an average path length of 450 m. Its angular deviation of 8.0° ensures smoother trajectories than traditional methods like Genetic Algorithm and Particle Swarm Optimization (PSO). Moreover, AMOPP achieves a 0% collision rate across all simulations, surpassing heuristic-based methods like Cuckoo Search and Bee Colony Optimization, which exhibit higher collision rates. Real-time responsiveness is another key strength of AMOPP, with an average re-planning time of 0.75 s, significantly outperforming A* and RRT*. The computational complexities of each algorithm are analyzed, with AMOPP exhibiting a time complexity of O(k·n) and a space complexity of O(n), ensuring scalability and efficiency for large-scale operations. The study also presents a comprehensive qualitative and quantitative comparison of 14 algorithms using 3D visualizations, highlighting their strengths, limitations, and suitable application scenarios. By integrating weighted optimization with penalty-based strategies and spline interpolation, AMOPP provides a robust solution for UAV path planning, particularly in scenarios requiring smooth navigation and adaptive re-planning. This work establishes AMOPP as a promising framework for real-time, efficient, and safe UAV operations in dynamic environments.

2 October 2024 in sinkron
Verified
Journal Article
Open access
PDF file
GAGregorius Airlangga
Gregorius Airlangga
In today’s competitive insurance market, accurately predicting customer interest in additional products, such as vehicle insurance, is crucial for optimizing marketing strategies and maximizing sales. This study presents a comparative analysis of three machine learning models such as XGBoost, RandomForest, and Logistic Regression to predict customer interest in vehicle insurance based on a dataset that includes demographic, vehicle, and policy-related features. The dataset was analyzed using five-fold cross-validation, and the performance of the models was evaluated using AUC-ROC, precision, recall, and F1-score. XGBoost demonstrated the highest recall (0.9525) and AUC-ROC (0.7854), making it the most effective model for identifying customers interested in vehicle insurance, though at the expense of lower precision (0.2585). RandomForest showed a more balanced trade-off between precision (0.3064) and recall (0.5341) but performed lower overall. Logistic Regression, while the most interpretable model, exhibited high variability in performance across different folds, with a lower average precision (0.2372). The findings of this research highlight that XGBoost is ideal for maximizing recall in high-volume campaigns, while RandomForest may be better suited for applications requiring fewer false positives. These results offer valuable insights into model selection based on business objectives and resource allocation.