This paper presents a method for estimating the ballistic coefficients (BC) of non-thrustered Low Earth Orbit (LEO) satellites using historical two-line element (TLE) data. The program developed in this study was implemented using the Python programming language, and detailed library specifications are provided in the paper. It was validated against the true ballistic coefficient values from 16 LEO objects in the High Accuracy Satellite Drag Model (HASDM) dataset, achieving overall absolute relative errors of less than 10%. The method employs the latest NRLMSIS 2.1 atmospheric model to compute atmospheric density. As a case study, it was applied to Indonesia's LAPAN-TUBSAT (A1), LAPAN A2, and LAPAN A3 satellites, yielding estimated BC values of 0.00114, 0.00825, and 0.00563 m²/kg, respectively. The estimation framework developed in this study can be broadly applied to other non-thrustered LEO satellites, where accurate BC determination is crucial for satellite dynamics studies and future space debris mitigation efforts.

This paper presents an observability analysis of a small Unmanned Aerial Vehicle (UAV) navigation system employing a Switching Extended Kalman Filter (SEKF) to evaluate its performance under various sensor availability conditions. SEKF fuses measurements from an Inertial Measurement Unit, Global Navigation Satellite System (GNSS), barometric altimeter, optical flow sensor, and attitude heading reference system, with the ability to adaptively switch modes based on sensor availability. Observability is assessed numerically using singular value decomposition of the observability matrix during different flight conditions, including straight and level-turn maneuvers. Results show that full sensor availability ensures complete state observability. In GNSS-denied scenarios, optical flow significantly contributes to horizontal velocity observability, while the barometer supports altitude correction. However, in the absence of both GNSS and optical flow, observability is reduced, particularly during steady straight flight. The analysis also demonstrates that flight dynamics, such as level turns, improve observability by exciting additional motion. These findings highlight the critical role of multi-sensor fusion and maneuvering in resilient UAV navigation and provide guidance for designing SEKF-based systems capable of maintaining performance in potentially degraded environments

Altitude-based flight state transitions are a fundamental part of autonomous aerospace systems, enabling precise execution of the mission sequences in a dynamic flight environment. This paper presents an in-depth analysis of an in-flight anomaly observed during the deployment of a dual-module CanSat "Ugrasena." The mission encountered an anomaly altitude reading from the barometric pressure sensor that triggered a premature payload separation. Detailed post-mission telemetry data analysis and ground-based sensor noise characterization revealed that transient pressure fluctuations induced an altitude change in inflight control logic. The false triggers cause an early separation of payload and compromise the data acquisition. To address this failure mode, mitigation strategies including advanced sensor filtering, redundancy through multi-sensor fusion, and adaptive threshold algorithms are discussed. The findings provide critical insights into the design of robust altitude-based control systems in small-scale aerospace platforms, ensuring resilience against environmental and sensor-induced disturbances.

Sixth-generation (6G) cellular networks aim to deliver diverse services with high mobility and flexible architectures, including the use of Unmanned Aerial Vehicles (UAVs). A key feature of 6G is device-to-device (D2D) communication, which enables direct links between user equipment (UE) without relying on base stations (gNB). To overcome the limitations of short-range D2D links, UAV-based relaying offers a promising solution. This paper reviews the integration of UAV relaying in D2D communication, focusing on Radio Resource Management (RRM) for UAV relaying in Search and Rescue (SAR) operations, and highlights key challenges and future research directions.

Tactical Unmanned Aerial Vehicles (TUAVs) with Vertical Take-Off and Landing (VTOL) play important roles in Intelligence, Surveillance, and Reconnaissance (ISR) missions. Therefore, reducing the visibility of VTOL TUAV from radar detection is a critical thing. This study aims to modify the aerodynamic characteristics and Radar Cross Section (RCS) of this kind of UAV. Two TUAV geometries were analyzed, namely non-VTOL and VTOL-modified. The VTOL modifications enable operations without a runway, enhancing flexibility in ISR missions. Simulation results show that adding VTOL features resulted in a 6.11% decrease in aerodynamic performance and an average increase in RCS of 17.79%. This reduction in aerodynamic performance is attributed to the complexity of airflow around the VTOL structure. At the same time, the increase in RCS is related to changes in the TUAV's shape and materials. In conclusion, this study has given valuable insights for developing more effective and efficient TUAVs in various ISR mission scenarios. It highlights the need for further research on shape and material optimization to reduce radar detection.

Meteorological conditions critically influence aviation safety and operational performance by affecting flight scheduling, routing, and hazard management. Unpredictable events such as wind shear, thunderstorms, and reduced visibility present challenges that compromise safety and efficiency despite advances in forecasting technology. This research presents an integrated analytical framework using the 2019 Airline Delays with Weather and Airport Details dataset. The proposed approach employs data preprocessing through spatio-temporal mapping and feature engineering, followed by predictive modelling using Artificial Neural Networks (ANN) and Extreme Gradient Boosting (XGBoost) to capture nonlinear relationships and identify influential variables. The combined ANN+XGBoost model achieves superior performance with a Mean Squared Error (MSE) of 50.12, Root Mean Square Error (RMSE) of 7.08, Mean Absolute Error (MAE) of 5.21, and Mean Absolute Percentage Error (MAPE) of 2.05, outperforming individual models. Delay pattern analysis reveals greater consistency under Instrument Flight Rules (IFR) conditions compared to higher variability in Visual Flight Rules (VFR) operations. The proposed framework offers a robust, data-driven decision-support tool for enhancing aviation risk management, operational planning, and resilience in varying weather scenarios.

This research presents the design and implementation of an FPGA-based data acquisition system for ground testing of satellite Payload Data Handling Systems (PDHS), with the LAPAN-A4 mission used as a case study. The proposed system replaces conventional X-band setups with a low-cost, laboratory-accessible solution by directly interfacing through an 8-channel LVDS at 25 MHz and streaming the data via USB 2.0 using the FX2LP controller. Experimental results demonstrated continuous acquisition with a stable throughput of 200 Mbps (8-bit mode) and up to 384 Mbps (16-bit mode), with nearly 100% valid frame detection after removal of residual idle data. Clock synchronization was managed by an FPGA PLL generating multiple domains, ensuring reliable buffering and streaming. FPGA resource utilization remained low (under 10%), confirming scalability for broader applications. Although detailed BER and jitter analysis could not be conducted due to equipment limitations, the achieved performance is sufficient for engineering validation, providing a practical and cost-effective alternative to traditional RF-based ground testing approaches.

e present the development of a highly-efficient 6x8 linear polarization array antenna for SAR sensors. The antenna operates at a center frequency of 5.3 GHz and is fabricated on NPC H220A substrate with a dielectric constant of 2.16 and a thickness of 1.6 mm. Side-lobe suppression is applied in the design, which enables the antenna to meet the stringent requirements for SAR systems with a range of over 1 km, including a gain of over 20 dB and a side-lobe level below -13dB. The proposed antenna design features a compact rectangular basic plane with a rectangular ring slot in the center, which allows for a high level of performance within a small footprint. Simulation results demonstrate a bandwidth return loss of 12% (5-5) and a gain of 21 dB, with a side lobe level exceeding -14 dB. The total dimensions of the antenna are 275 mm x 380 mm. The antenna is fabricated using advanced double-substrate technology with NPC-H22A material, which has a dielectric constant of 2.17 and a thickness of 1.6 mm. This design produces highly effective linear polarization, making it an ideal solution for SAR applications

Satellites consist of a bus system and a payload system, with the Electrical Power System (EPS) as one of the main bus components. Within the EPS, the Power Distribution Unit (PDU) is critical for distributing and regulating power to all subsystems at different voltage levels, making its reliability essential for mission success. This study evaluates the reliability of a buck converter in the satellite PDU using the MIL-HDBK-217F standard and mission scenarios. Reliability modeling was conducted using the Markov model and Weibull distribution to estimate the reliability function (R(t)) and Mean Time To Failure (MTTF). The base failure rates of components-inductor (0.002), low-ESR capacitor (0.01), diode (0.005), and switching MOSFET (0.008) failures per 10⁶ hours- were corrected by applying temperature, quality, environmental, and power stress factors. Simulation results show that the 28 V rail has the lowest MTTF of about 1022 years, while the 3.3 V rail achieves the highest at around 1986 years. The power stress factor (πS) is identified as the dominant variable influencing lifetime. The Weibull approach proves suitable for mission-based satellite applications and provides a useful reference for designing more reliable PDUs and predicting system lifetime more accurately.

High-speed image data acquisition on low-cost FPGAs is often constrained by timing closure, inter-lane skew, and inconsistent reporting, which collectively make sustaining rates above 150 Mb/s difficult in remote-sensing payloads. This paper presents a practical framework to design high-speed imaging data acquisition systems on low-cost FPGAs, structured around requirement modeling, a modular receiver architecture, and a reproducible verification protocol. The architecture features 66 MHz PLL-based clock conditioning, multi-channel serial-in/parallel-out conversion, dual-marker framing with SYNC1 (0000h) and SYNC2 (3FFFh), and inter-bank phase alignment to guarantee deterministic timing. Validation on a Cyclone IV shows full-stream throughput R_full of 178.282 Mb/s (incl. idle) and active throughput R_active of 184.08 Mb/s (valid only), with clock deviation of 0.24% at 66 MHz; T_frame of 2.271629 ms (rounded to the 149,857-cycle budget to reflect expected drift); 100% marker detection; and no frame loss observed over 60-minute continuous runs. These findings demonstrate that the proposed approach supports reliable image acquisition more than 150 Mb/s (validated at 178.282/184.08 Mb/s) on cost-constrained FPGA platforms, providing a standardized foundation for reporting and replication across teams.

High-speed data acquisition with precise synchronization and integrity is critical for Payload Data Handling Systems (PDHS) in Low Earth Orbit (LEO) satellites such as LAPAN-A4, particularly during ground testing. Conventional approaches remain inefficient as they rely on antennas and receivers, making it difficult to efficiently handle high-rate PDHS data and payload streams such as camera outputs under different test configurations. To address these challenges, this study presents an FPGA-based acquisition system that integrates Low-Voltage Differential Signaling (LVDS) and is validated through VHDL simulation. The proposed design enables efficient reconstruction and verification of CCSDS packets and camera data, significantly improving testing flexibility. System performance was evaluated in terms of bandwidth utilization, repetition time, and sustained throughput. Results demonstrate a stable PDHS data rate of 200 Mb/s and an active camera frame duration of 2.191513 ms, closely matching the specified 2.191 ms. These outcomes confirm the system's reliability, consistency with mission requirements, and scalability for higher data volumes or future satellite configurations. Overall, the system provides an effective and practical solution for LAPAN-A4 and upcoming LEO satellite missions requiring robust ground testing of high-speed data acquisition.

This study investigates the influence of orientation and stiffener design (solid, isogrid, orthogrid, and honeycomb) on the natural frequency of a 6U CubeSat deployable solar panel (DSP). Using finite element modal analysis, three scenarios were evaluated: constant stiffener thickness, constant DSP mass, and constant first natural frequency. Results show that with equal thickness, the orthogrid design offers the most efficient mass reduction with minimal frequency loss, while the orthogrid rotated 45° performs worst. When mass is held constant, honeycomb stiffeners yield the highest natural frequency increase. Under equal frequency conditions, honeycomb stiffeners achieve the greatest mass reduction, up to 29.1% lighter than the solid type. These findings provide guidance for selecting stiffener configurations that optimize mass efficiency, frequency performance, or a balanced trade-off to meet CubeSat launch constraints.

This study proposes and evaluates a novel 12U CubeSat deployer design featuring external Wire Rope Isolators (WRIs) to mitigate severe launch-induced vibrations. Finite Element Analysis, comprising modal and random vibration analyses, was systematically performed on three distinct WRI configurations. The objective was to identify the most effective arrangement for optimal vibration isolation. Results from modal analysis showed varying dynamic characteristics among configurations, with Variation 3 exhibiting slightly lower natural frequencies in lower modes, indicating reduced stiffness. Crucially, random vibration analysis revealed that Variation 3, utilizing 12 WRIs in compression and 4 in shear, consistently achieved the lowest peak acceleration responses across all sensors and axes. These findings definitively establish Variation 3 as the optimal configuration for superior vibration attenuation, essential for safeguarding satellite integrity and ensuring successful deployment in harsh launch environments.

This study investigates the implementation and performance of scheduled off-nadir imaging using Long-Time Telemetry (LTT) prediction on LAPAN-A2 satellite case. Scheduled imaging is attempted due to high operator effort and unreliable radio frequency links in real-time off-nadir imaging, it also enables image acquisition beyond ground station coverage. The method utilizes inertial pointing mode with three reaction wheels to control satellite attitude (roll, pitch, yaw) based on input angle commands. Roll angle is determined by the target's distance from satellite ground track, while pitch angle depends on the target's latitude. The prediction combines two previous LTT logs which each them contains satellite's health information for the last 86 minutes to estimate satellite's attitude required for observation, using trend-based extension to improve the accuracy. Image acquisition results demonstrate pointing shift for about 12.2 km for a target with 1.7° roll angle and 7.15 km offset for 9.4° roll angle. Compared to previous years that using real-time off-nadir imaging which result to successful imaging during 2020 Krakatau eruption, recent scheduled off-nadir performance shows inadequate certainty. Real-time imaging benefits from closed-loop nature, allowing interactive attitude correction, while scheduled imaging may suffer from attitude drift and LTT interpretation inaccuracies. These challenges highlight the need for more accurate predictive models to support scheduled imaging. Wheel speed analysis on Y axis shows the off-nadir cases indicating that combining Y and Z wheel speeds, rather than relying on angle commands alone, more effectively reduces pointing offsets in subsequent off-nadir acquisitions.

The LAPAN-A3 satellite employs a momentum-bias attitude control system using a reaction wheel and magnetorquers. By using momentum bias, maintaining the angular momentum vector is important. However, external disturbance torques can reduce angular momentum, making momentum biasing necessary. Telemetry data from 2022 to 2025 were analyzed through linear modeling with the least squares method and validated against the actual observed change of angular momentum. The results indicate a consistent linear relationship between the change in angular momentum and the activation duration of Coil Z. Modeling of ascending passes produced stable gradient values (2.52-2.53) during the period from 2022 to 2024, confirming the effectiveness of Coil Z as an external torque. Validation showed that most errors were within the 0-11% range, with the general equation yielding a lower Root Mean Square Error (RMSE) (38.9) compared to the annual equation (41.3). The linear approach proved sufficiently accurate in representing the relationship between the change in angular momentum and Coil Z activation duration, particularly on ascending passes, and provides a basis for developing automated attitude control strategies for low Earth orbit satellites. The results also indicate that the satellite's magnetic characteristics undergo negligible changes with aging.

Soil moisture is an integral variable of hydrological processes that largely influences agricultural productivity and environmental sustainability. Motivated by the scarcity of research on soil moisture monitoring in Bangladesh, our study lays the foundation for soil moisture regression research within its diversified agricultural landscapes, spanning 30 Agro-Ecological Zones. We developed a cost-efficient data pipeline that compiled a weekly dataset (2022-2024) incorporating dynamic (e.g., vegetation and moisture indices, temperature, rainfall) and static (e.g., elevation, landcover) features by procuring open-source data from SMAP, MODIS, Sentinel-1, CHIRPS, ERA5 via Google Earth Engine. Utilizing the acquired dataset, we evaluated Linear Regression, Random Forest, Support Vector Regression, LightGBM, and a shallow Artificial Neural Network for soil moisture estimation. For week-ahead forecasting, we implemented LightGBM and our reduced-feature Random Forest. Feature screening via Pearson correlation and SHAP value analyses identified evapotranspiration, rainfall, moisture indices, and elevation as dominant predictors. LightGBM achieved the highest performance for concurrent estimation (R² 0.85 ± 0.01, RMSE 0.041 ± 0.001) and also excelled in week-ahead forecasting (R² 0.78 ± 0.01, RMSE 0.050 ± 0.003). We faced limitations concerning data gaps and absence of in-situ validation that constrained fine-scale applicability. Future work should integrate in-situ data, employ advanced gap-filling techniques, and explore downscaling for augmenting precision farming in Bangladesh and similar regions of Asia.

This paper proposes an integrated strategy to analyse the progress of selected United Nations Sustainable Development Goals (SDGs 1, 2, and 13) using Earth Observation (EO) data and deep learning (DL) based classification. The study examined Mauritius's Black River district's settlement patterns, vegetation, bare land, and forest areas using satellite images from the Satellogic constellation. Preprocessing, categorization, and SDG-specific geographical mapping were all carried out. The proposed DL model achieved up to 98% overall accuracy.

Remote sensing imagery is frequently degraded by dense haze and thin clouds, which hinder accurate Earth observa- tion and downstream analysis. In this paper, we propose a densely connected U-Net-based deep learning architecture tailored for haze and cloud removal in satellite images. The proposed model is trained using available RS-Haze datasets and transfer learning was performed on Resourcesat-2-LISS3, Cartosat-2S MX and Indian Satellite data. The model outcome was evaluated using a combination of SSIM and MSE loss to preserve structural integrity and pixel-level detail. After training for 250 epochs, the model demonstrates strong generalization across varying atmospheric conditions, effectively removing dense haze while preserving critical land features such as river boundaries, urban layouts, and agricultural zones. Quantitative results confirm improvements in PSNR and SSIM over existing baselines, and qualitative assessments further validate the model's capability to enhance image clarity for remote sensing applications. The proposed approach offers a promising tool for improving the image clarity for geospatial analytics in hazy environments.

Drought has become an increasingly serious threat in Indonesia over recent decades, expanding in both coverage and duration, and significantly impacting agriculture, water availability, and food security. Although Indonesia typically experiences year-round rainfall as a tropical country, elevated Tropical Pacific Ocean surface temperatures can trigger El Niño events, reducing rainfall and extending drought periods across Indonesian territories. This study aims to map the spatial distribution of drought potential on Java Island using satellite data and machine learning techniques. The research employs meteorological, agronomic, and hydrological drought indices: Keetch-Byram Drought Index (KBDI) for atmospheric conditions, Vegetation Health Index (VHI) for biological conditions, and Topographic Wetness Index (TWI) for hydrological conditions. Population density data were incorporated to represent socioeconomic aspects across Java Island. The random forest method was used to distinguish between dry and normal conditions. Results were validated using drought data from the National Disaster Management Agency (BNPB). Classification results for July 2019 showed (53.5%) of dry land, with a validation accuracy of 84.5%. This research can serve as a reference for developing early warning systems for drought disasters in Java.

This study presents a federated learning approach utilizing NDVI (Normalized Difference Vegetation Index) data to assess plant health across Turkey. The data were acquired from the USGS database via the Landsat 8 satellite, using the Red and NIR bands. Following acquisition, the data were preprocessed by segmenting them into patches and assigning corresponding labels. These were then used to train Convolutional Neural Network (CNN) models. Training initially involved three CNN architectures: Vanilla CNN, ResNet50, and EfficientNetB0, with region-specific models developed for each. However, EfficientNetB0 was excluded from later stages of the study due to its inadequate performance. Vanilla CNN and ResNet50 architectures, which successfully produced regional models, were incorporated into the federated learning framework owing to their training efficiency and privacy-preserving advantages. As a result, two distinct federated learning models were developed and tested using previously unseen data. Standard performance evaluation metrics commonly used in CNN-based studies were employed for model assessment. Although the federated learning model based on the ResNet50 architecture achieved high accuracy, it suffered from issues related to class imbalance and prediction instability, and was thus deemed unsatisfactory. In contrast, the Vanilla CNN-based federated learning model demonstrated high accuracy, low error rates, and balanced classification performance, marking it as a successful model. In conclusion, the study produced an effective federated learning model capable of detecting vegetation health status across Turkey.

In this work, we propose a new image preprocessing technique for more accurate tree species classification, named Pseudo Tree Crown (PTC). We investigated the effectiveness of the PTC preprocessing method in improving the classification accuracy of individual tree species (ITS) using a high-resolution unmanned aerial vehicle (UAV) remote sensing images covering the North America and based on the classification accuracy obtained before and after image preprocessing. It was utilized the collection of tree species on the Roboflow platform, including fir, spruce, pine, and trembling aspen, and analyzed using several machine learning (ML) and deep learning (DL) models, including YOLOv10, YOLOv11, PyTorch ResNet50, and Random Forest, to evaluate classification performance. The experimental results showed that the data obtained through PTC preprocessing achieved excellent performance across classification models, improving both accuracy and recall for multiple tree species. Specifically, the performance is exceptionally well in species with stable structural patterns.

Despite UAV-photogrammetry being a potential solution for data acquisition, the mapping process cannot always be conducted automatically due to various factors. Therefore, it is imperative to develop automation methods that can be effectively applied to detailed-scale mapping. This study investigates the effectiveness of utilizing airborne imageries as training datasets for building segmentation in UAV imagery using deep learning. In addition to predicting the original UAV imagery, we enhanced accuracy by employing histogram modification. The results of our study demonstrated that this approach effectively improved the quality of building segmentation results obtained from UAV imagery. Initially, the original UAV imagery yielded an accuracy score of 69.51% and an IoU score of 71.50%. However, after implementing histogram modification, the accuracy score increased to 73.18%, while the IoU score experienced a more significant improvement, reaching 80.52%. These findings indicate that improving deep learning results is not only about increasing training data or refining model design; modifying test data also plays an important role in achieving better performance.

Accurate ground-space classification in sensor horizon is fundamental for autonomous navigation, horizon detection, and Earth observation applications. This paper presents a novel binary classification approach that combines enhanced Convolutional Neural Networks (CNN) with adaptive border removal and multi-modal analysis for distinguishing ground-dominant from space-dominant regions in LAPAN TUBSAT satellite imagery. Our method addresses critical challenges, including vertical horizon detection, border artifacts, and class imbalance, through a comprehensive framework incorporating brightness analysis, spatial clustering, and confidence-weighted decision fusion. Experimental results on 2,503 satellite images demonstrate superior performance with 97% accuracy, effectively handling edge cases where traditional methods fail. The proposed system successfully identifies space regions with dark pixel ratios above 0.4 and adapts to both horizontal and vertical horizon configurations, making it suitable for real-time satellite applications.

The increasing volume of data generated by Satellite-based Automatic Identification Systems (S-AIS) demands ground segment architectures capable of transforming raw satellite telemetry into timely, actionable maritime intelligence. Existing systems often rely on passive data handling and lack integrated analytical capabilities, limiting their effectiveness for real-time maritime domain awareness (MDA). This paper presents an integrated ground segment framework designed for the upcoming NEO-1 small satellite mission. The framework comprises three core subsystems: (i) an Automated Data Pipeline (ADP) for end-to-end data processing, including CRC-based validation, ITU-compliant decoding, and multi-stage cleaning; (ii) an Intelligent Analysis Module (IAM) employing interpolation, clustering, classification, and forecasting to extract operational insights; and (iii) a Dissemination System providing secure access via a REST API and web-based visualization. The modular, containerized architecture ensures scalability, portability, and real-time performance. This work contributes a transition from data-passive to intelligence-active ground segments, offering a scalable blueprint for next-generation S-AIS missions supporting global maritime security and surveillance.

Super-resolution of satellite images plays a crucial role in enhancing the quality and utility of remote sensing data for various applications including environmental monitoring, urban planning, and disaster management. This paper presents a supervised approach for satellite image super-resolution using Vision Transformers, specifically leveraging the Swin2SR model with window-based attention mechanisms. Our methodology creates low-resolution datasets by patch extraction and 2× downsampling from high-resolution satellite images of Indian city and surrounding sub-urban region obtained from the ISRO platform, enabling supervised training with paired LR-HR datasets generated from a single source image. We implement a custom Swin2SR architecture with cyclic learning rate scheduling using the triangular-2 format to optimize training efficiency. Experimental results demonstrate substantial improvements over traditional approaches, achieving a PSNR of 33.85 dB, SSIM of 0.9673, MSE loss of 0.0043, and training loss reduction to 8.2. These results validate the effectiveness of Vision Transformers for supervised satellite image super-resolution tasks using single-image patch-based training with simulated LR-HR pairs.

Peatlands play a crucial role in maintaining global climate balance due to their ability to store large amounts of carbon and act as a sink for greenhouse gas emissions such as methane and carbon dioxide. However, peatland ecosystems have been increasingly degraded due to deforestation, infrastructure development, and land conversion, leading to land subsidence and higher CO2 emissions. This issue leads to disaster risks such as ecosystem disturbance, increased flood risk, and infrastructure damage. Central Kalimantan Province, particularly in the Palangkaraya area, is among the regions with extensive peatland areas affected by this phenomenon. This study aims to analyze the pattern and velocity of vertical deformation using the Persistent Scatterer-Distributed Scatterer Interferometric Synthetic Aperture Radar (PSDS-InSAR) approach, based on Sentinel-1 and ALOS PALSAR-2 imagery acquired between 2017 to 2022. This method combines the advantages of Persistent Scatterer (PS) techniques for urban areas and Distributed Scatterer (DS) techniques for vegetated areas to increase deformation pixel density, thereby yielding broader spatial coverage. Data processing was performed using software from the ISCE Framework, with results validated against groundwater level data. Land subsidence observed in the Palangkaraya peatland area aligns with the hydro-climatological conditions of the region. A distinct pattern of land subsidence is evident during the dry season, accompanied by a decline in groundwater levels in the next few days. The velocity of ground deformation varies from −2.24 to 2.36 cm/year for Sentinel-1 and −1.31 to 1.08 cm/year for ALOS PALSAR-2, with RMSE values of 0.4279 cm/year and 0.6351 cm/year, respectively.

Accurate and timely mapping of rice cultivation is essential for food security monitoring, water resource management, and agricultural planning, particularly in monsoon-dependent regions like Tamil Nadu. This study presents a seasonal rice area mapping framework using dense Sentinel-1 VH-polarized Synthetic Aperture Radar (SAR) time series and supervised machine learning on the Google Earth Engine platform. Focusing on Thanjavur District, a significant part of Cauvery delta zone, Tamil Nadu, the methodology captures crop phenological progression during two major rice-growing seasons, pre-monsoon crop cultivation period and peak monsoon crop cultivation period. VH composites were generated and stacked to form multi-band datasets representing crop development. Field-verified samples of rice and non-rice classes were used to train Random Forest (RF) and Classification and Regression Tree (CART) classifiers. The RF model outperformed CART across both seasons, achieving overall accuracies of 95.44% (Kuruvai) and 94.70% (Samba), with corresponding Kappa coefficients of 0.9085 and 0.8940. CART showed lower performance particularly during the Samba season. The results demonstrate the potential of integrating SAR time-series with ensemble machine learning for operational, seasonally disaggregated rice mapping in cloud-prone environments. The workflow, fully implemented in GEE, is scalable and transferable for regional agricultural monitoring across rice-dominant landscapes.

This study improves the quality of NDVI data for multitemporal vegetation monitoring in the Bedadung Watershed (DAS), Jember, through the application of the Discrete Wavelet Transform (DWT) filter based on Daubechies 4 (db4). The analysis stages include masking the study area, calculating NDVI from the red and NIR bands, and applying level 1 DWT filtering to remove high-frequency components. Evaluation was carried out using statistical metrics such as RMSE, MAE, the coefficient of determination, mean NDVI, and visualization of NDVI value distribution. The results show that DWT is effective in reducing noise without altering the spatial structure of the data. The coefficient of determination reached 0.89, with low RMSE and MAE values (0.05 and 0.01), indicating high accuracy. The average NDVI increased from 0.1091 to 0.1147 in the 2019-2023 period, indicating improved data quality. NDVI distribution also became more uniform after filtering. These findings support the use of DWT db4 as a reliable method for long-term vegetation analysis, particularly in supporting land conservation efforts and flood mitigation in the Bedadung Watershed

Flooding is one of the most severe environmental disasters, impacting millions of lives and causing significant economic losses. Effective flood monitoring is therefore essential to support timely disaster response, resource management, and damage assessment. However, current optical remote sensing methods to monitor floods, particularly multispectral imaging (MSI), are limited by low spectral resolution and rely heavily on complex post-processing algorithms to extract accurate flood information, leading to higher computational costs and reduced efficiency. To address this limitation, this study presents the hardware design and simulation of a high-resolution hyperspectral lens system optimized for CubeSat deployment. The optical system was modeled and optimized in ANSYS Zemax OpticStudio, targeting a spatial resolution of 30 meters with a 20-degree field of view across 24 spectral bands from 465-810 nm. The final design achieved an effective focal length of 133.394 mm and a total system length of 142.087 mm, with a compact weight of approximately 590 grams, making it suitable for 2U CubeSat integration. Performance analysis confirmed strong imaging quality, with an Airy disk radius of 7.937 815 micrometers, Strehl ratio of 0.942, MTF contrast of 0.4 at 34 cycles/mm, and significant reduction of coma (96%), astigmatism (101%), and field curvature (98%). These results demonstrate the system's capability to deliver high-resolution imaging with minimal reliance on post-processing, offering a compact and efficient optical solution for spaceborne flood monitoring.

Sumatra Island, one of the largest forest-covered regions in Indonesia, has experienced substantial forest loss over the past decade. This study investigates the spatio-temporal of forest degradation and its impact on biomass carbon from 2010 to 2020. We utilised openly accessible remote sensing datasets integrated with GIS-based spatial analysis to map forest cover changes and identify degradation patterns caused by deforestation, the increase in cropland coverage, and the continued spread of oil palm plantations. In addition, we employed the Random Forest regression algorithm to model forest biomass carbon, encompassing both aboveground biomass carbon (AGBC) and belowground biomass carbon (BGBC) across ten provinces in Sumatra. The results revealed a substantial reduction in natural forest area during the observation period, totalling 12.07 million hectares (Mha), with the highest losses recorded in South Sumatra, Riau, and North Sumatra. Quantification of biomass carbon showed varying results in relation to forest area loss during the observation period. These findings underscore the urgent need to strengthen forest conservation policies and spatial land governance to support Indonesia's climate mitigation commitments under the Enhanced Nationally Determined Contribution (ENDC) and the goal of achieving net-zero emissions by 2060 or earlier.

Research on bathymetric LiDAR may advance and be used to map coastal areas and shallow waters. The simplest method used in a bathymetric LiDAR system is the time-of-flight. In this work, we constructed a biaxial indirect time-of-flight (iToF) LiDAR system utilizing a diode laser with a wavelength of 405 nm. The 405 nm iToF LiDAR system was validated by measuring object distance in free space and aquatic environments throughout the 1.93-6.00 m measuring range. The analysis of the measurement results indicates that the error value varies between 2.82% and 13.26% when compared to the reference distance. Validation findings confirm the 405 nm iToF LiDAR biaxial system's reliability for shallow water bathymetry applications.

Rapid environmental shifts driven by human activity and climate change, Marine Protected Areas (MPAs) are essential tools for biodiversity conservation. However, their effectiveness is increasingly challenged by climate-related disasters like marine heatwaves (MHWs) and algal blooms. This research evaluates these compounding threats to MPAs in the border seas of Indonesia and Malaysia by analysing MHWs and Coastal Eutrophication Potential (CEP). Utilizing 40 years of sea surface temperature (SST) data from OSTIA and 9 years of chlorophyll-a data from OLCI, the study performs a long-term spatial and temporal analysis. The findings reveal that the Malacca Strait consistently exhibits high CEP, which intensifies within its MPAs during the northwest monsoon due to heightened nutrient runoff from heavy rainfall. In contrast, the Natuna Sea maintains low eutrophication potential. The region's MHWs are more strongly correlated with El Niño-Southern Oscillation (ENSO) events than the Indian Ocean Dipole (IOD). Critically, a major MHW event from November 2023 to November 2024 highlighted that MPAs are vulnerable to the simultaneous, compounding impacts of extreme heat and severe algal blooms. These dual threats underscore the urgent need for robust bilateral strategies between Indonesia and Malaysia. Both nations must prioritize integrated, research-driven management to develop a more effective and resilient strategy for protecting their vital MPAs in the face of escalating climate change.

Bayer encoded images are natively captured by most satellite sensors, yet existing detection models typically require full RGB conversion before inference. This conversion inflates memory usage by a factor of three and can degrade image-specific structural patterns. In this work, we propose a Bayer-Aware Convolutional (BAC) layer that adapts the first convolutional layer in YOLOv11 to exploit the raw Bayer pattern directly, without RGB conversion. Specifically, we encode prior knowledge of the BGGR Bayer mosaic into the convolution weights, further enhanced by a channel-wise attention mecha- nism. This modification allows the network to better leverage inherent spatial correlations in Bayer-encoded inputs, leading to improved accuracy and reduced storage requirements. We demonstrate the effectiveness of our method on oriented ship detection in satellite imagery, achieving up to a nearly 1% mAP improvement over conventional approaches while reducing input storage requirements by 3 times. This enables more efficient on- device and satellite-side inference, opening new directions for resource-aware remote sensing applications

Target detection in sea clutter for satellite-based images is studied. The detection process is based on the Constant False Alarm Rate (CFAR) algorithm. To obtain the adaptive threshold, we conduct a thorough spatial statistical analysis of the sea clutter. The common issue with Satellite-based Synthetic Aperture Radar (SAR) images is the contamination with speckle noise. The goal of the paper is to study the effect of the speckle noise on the statistical properties of the sea clutter. Based on the experimental data gathered from the Canadian RADARSAT-1 satellite, we demonstrate that the Weibull, Rayleigh, and K distributions are capable of modeling the statistical properties of the sea clutter in the presence of the speckle noise more precisely while Weibull, Gamma, inverse Gaussian, and Log-normal distributions describe the statistical properties of the sea clutter with higher accuracy when the speckle noise is removed. The goodness-of-fit measure is based on the Kullback-Leibler (KL) divergence metric. The speckle noise removal process is based on median filtering with the Peak Signal to Noise Ratio (PSNR) of the image as a measure for the filter parameter estimation.

The presented results, indicate that the Weibull distribution is able to model the statistical properties of the sea clutter both in the presence and absence of the speckle noise with high accuracy. To estimate the parameters of the Weibull distribution, we propose a method based on the Mellin transform which compared to the existing techniques, provides a closed-form and untangled solutions for both parameters.

Carbon Monoxide (CO) is one of the air pollutant gases. CO distribution can be analyzed using geospatial information to prevent its increasing concentration, includes in several cities region in East Java. This study aimed to assess CO distribution and its contributor by utilizing two satellite data Sentinel-5P and Landsat 8 OLI/TIRS. Temporal satellite images was processed using Google Earth Engine. This research was conducted in Surabaya city, Mojokerto city, and Mojokerto regency from 2018 to 2023. These cities have different characteristics so that the CO concentration produced would also be different. Pearson correlation was conducted to determine the relationship between CO extracted from Sentinel-5P and Landsat-8, including social factors, i.e population density, urban factors of building distribution density and transportation factors. Urban factors was analyzed using NDBI and NDVI, while transportation factors was analyzed from the total number of motor vehicles. Results showed correlation coefficient varied for each city. For social factors, the highest correlation was in Mojokerto district approximately 0.7329 in 2023. This indicates increases population density affects CO concentration. The urban pattern of high building density in Surabaya city developed using NDBI contributed to the highest correlation with CO concentration approximately 0.6868 in 2022. This indicates that building density affects CO concentration. Transportation factors produced the highest correlation in Mojokerto city (0.6774) that affects increases CO concentration. While high vegetation density mapped using NDVI had the highest negative correlation in Mojokerto regency in 2018 with approximately 0.5706. This showed that the extensive vegetation density could reduce CO concentration.

Rapid expansion of remote sensing (RS) image archives necessitates the development of advanced content-based image retrieval (CBIR) methods to effectively organize and access high-dimensional visual data. Deep metric learning, particularly with triplet loss, has shown promising results in retrieval tasks by embedding semantically similar images closer in feature space. However, conventional fixed-margin triplet loss functions are inadequate for multi-label RS archives, where high intra-class variability and inter-class similarity is observed. These limitations often result in ineffective training due to the dominance of trivial triplets and the zero-loss problem. In this work, we propose a novel multi-label adaptive triplet learning approach for CBIR in RS archives. The proposed method introduces an adaptive margin approach that dynamically adjusts the triplet loss margin based on the hardness of each sample, computed as the difference between anchor-positive and anchor-negative distances. This adaptive strategy enables the model to focus on informative (hard) triplets and suppresses the influence of trivial ones, leading to more discriminative feature embeddings. We evaluate our proposed approach, on two challenging multi-label RS archives: MLRSNet and UCMerced. Experimental results demonstrate that the proposed method consistently outperforms fixed-margin triplet loss and several state-of-the-art CBIR approaches, particularly in retrieval accuracy and robustness. These findings confirm the effectiveness of adaptive margin learning in handling the complexities of multi-label RS image retrieval.

The progressive urban expansion in highly urbanized cities of the Philippines, such as Butuan City, necessitates a robust supply base and access to essential services to achieve sustainable development in environmental, social, and economic dimensions. This paper investigates population dynamics across barangays from 2010 to 2020, revealing trends of both growth and decline, with implications for urban planning and resource allocation. Additionally, the study evaluates land cover changes between 2010, 2015, 2020, and 2023, highlighting the increase in built-up areas by 510.35%, and a corresponding decline in farmlands and forests. Using advanced techniques like Geographic Information Systems (GIS) and remote sensing, the research correlates land-use changes with socio-economic factors, such as population growth, urban development, and economic shifts. The results underscore the complexities of urban growth, linking it to both opportunities and challenges for sustainable urban governance. The findings emphasize the need for continuous monitoring and adaptation to manage Butuan City's evolving urban landscape and to foster long-term resilience for its communities.

Multispectral images acquired by optical remote sensing satellite in tropical areas, like Indonesia, are frequently affected by some distortions as a result of the presence of clouds and thin clouds or haziness. These distortions cause a decrease in image quality and the accuracy of image interpretation. Researchers have developed various methodologies to detect and remove hazed images achieved by optical remote sensing satellite. Nevertheless, the method development of haze elimination remains a challenge in improving the good aspect of multispectral optical remote sensing images, especially in providing cloud-free mosaic images. In this research, a haze removal technique on SPOT 6 and SPOT 7 images using a haze index algorithm was introduced. The algorithm is based on the reflectance values of the blue and red channels on SPOT 6 and SPOT 7 images. This paper tests the correlation of hazy images with a clean reference image to look for the most optimal haze factor coefficient value to remove haze in SPOT 6 and SPOT 7 images. In this haze removal process, a haze index algorithm and predetermined haze limit values were applied. Based on empirical discovery in the research, it is concluded that this correlation test obtains a haze factor coefficient that can provide optimal results in the haze removal process using the haze index algorithm for SPOT 6 and SPOT 7 images.

This paper proposes a direction finding system for detecting drones or unmanned aerial vehicles (UAVs) based on radio signals. The proposed direction finding system consists of four uniform circular array (UCA) elements connected to a single receiver unit via a single pole n throw switch (SPNT). The receiver connects to each antenna element subsequently. The direction of signal arrival is determined using ensemble machine learning directly through the RF signal without the need for baseband processing such as the demodulation. Experimental results show that the proposed system has good performance and also has low complexity.

The MSSR system plays a critical role in the aviation sector for precise aircraft identification and tracking, operating at designated frequencies of 1030 megahertz and 1090 megahertz. This research presents the design and evaluation of a reconfigurable microstrip antenna that utilises a passive tuning element, specifically a capacitor, to enable dual-frequency operation. The antenna was designed using CST Studio Suite 2019, followed by fabrication and experimental validation. The integration of the tuning capacitor enables a frequency shift from 1090 megahertz to 1030 MHz. At 1030 megahertz, the return loss achieves values of -27.60 dB in simulation and -17.78 dB in measurement, with VSWR values of 1.16 in simulation and 1.05 in measurement. At 1090 megahertz, the return loss reaches -25.36 dB in simulation and -13.35dB in measurement, with VSWR values of 1.13 and, 1.08 respectively. The measured gain is 1.13 decibel isotropic at 1090 MHZ and 1.20 decibel isotropic at 1030 MHz. These results demonstrate that the proposed antenna design with passive capacitive tuning offers effective frequency reconfigurability while maintaining stable performance, making it suitable for MSSR applications.

This paper presents a system engineering modeling approach for the development of a CubeSat-based Automatic Packet Reporting System (APRS) using the V-Model lifecycle to support emergency communication. Using this Model as a widely adopted engineering methodology that emphasizes verification and validation (V&V), illustrates how structured systems development ensures mission reliability and performance. Through a detailed analysis of system requirements, design, implementation, integration, and testing phases, it demonstrates a systematic road map to realize Cubesat missions for emergency communication. The APRS payload enables real-time position and message relay, supported by simulations, subsystem testing, and verification processes, similar with Assembly, Integration, and Testing (AIT) workflows. The results validate the relevance of the V-Model for nano satellite programs with emergency communication objectives.

This paper comprehensively analyzes the communications link performance of RIDU-Sat 1, Indonesia's first collaborative nano satellite developed by the Republic of Indonesia Defense University (RIDU) with an APRS payload for educational and research purposes. The satellite operates in the VHF frequency band at 145.925 MHz using GMSK modulation at 1200 baud for its Telemetry, Tracking, and Command (TT&C) system. Designed with a minimal ground station configuration comprising a software-defined radio (SDR), a Yagi antenna, and open-source software tools such as GNU Radio and YAMCS, this study evaluates whether such a low-cost setup can reliably support continuous satellite-ground communication. The research involves detailed link budget calculations and Doppler shift characterization for both uplink and downlink paths. Uplink results indicate that even at a low elevation angle of 0°, the system maintains a positive link margin of 0.54 dB, which increases to 19.63 dB at 90°. In comparison, downlink link margins range from 8.24 dB at 0° to 22.33 dB at zenith. Doppler analysis reveals a maximum frequency deviation of ±3424 Hz, which must be addressed during signal acquisition and tracking. The findings prove that reliable communication can be achieved using low-budget commercial off-the-shelf (COTS) components, supporting the feasibility of nanosatellite development in academic settings and promoting the use of accessible amateur radio infrastructure for space education in emerging countries.

Lately small satellites number rapidly increasing along with their capability to do complex missions. Momentum-bias attitude control, with its simplicity and stability, found its place on small satellites. However, due to their high momentum, this control strategy has low agility. Agility is a feature that will enhance the satellite capabilities, especially for Earth observation satellites. Agility in the form of slew maneuver can help the satellite to observe any areas not exactly under its orbit. Slewing a high-momentum satellite is a challenging process. A momentum transfer, a process of exchanging angular momentum between the actuators and the platform without any external torque, is proposed to handle the slew maneuver in this research. A combination of trajectory planning and tracking control is used. By using this strategy, only 2 wheels are needed, and no attitude information is needed for the control loop. The controller is analyzed through a numerical simulation using the LAPAN-A2 parameter, a 87 kg small satellite orbiting at 600 km altitude. The research found that the trajectory planning is crucial, as it needs to take into account the whole dynamics of the satellite and actuators, and be aware of the actuators' constraints. This research shows that a well-planned trajectory of momentum transfer can achieve a steady state error of 0.21°, 0.46°, and 0.37° in roll, pitch, and yaw under low Earth orbit disturbances.

Technology Advancement in satellite provides a wide range of applications, including broadcasting, meteorology, remote sensing, and amateur radio. Among these, the Es'hail-2/QO-100 satellite stands out as the first geostationary satellite dedicated to amateur radio communication. It was built by Mitsubishi Electric and launched aboard a SpaceX Falcon 9 rocket carrying two transponders specifically designed for HAM radio use. While the station of the Es'hail-2/QO-100 satellite is commonly established in most countries, for Cambodia, it is the first station; hence, in this paper, we will indicate the design and implementation of a ground station in Cambodia, which enables two-way communication with amateur radio operators across more than 100 countries. The station setup includes a UHF transceiver, Software Defined Radio (SDR), parabolic reflector antenna, dual patch antenna, frequency upconverter, downconverter, and power amplifier. The configuration successfully demonstrates the global reach of amateur satellite communication from Cambodia, including a milestone first contact with Antarctica. Communication latency remains low, between 240-300 milliseconds, allowing smooth and real-time voice communication using SSB. The system also shows high reliability, particularly during dry seasonal conditions.

A C-Band frequency-modulated continuous wave (FMCW) radar device is designed and developed for the aim of detecting, imaging and classifying human and/or animal micro-Doppler movements. The implemented prototype can detect micro-Doppler signatures up to ±500 Hz. A micro-Doppler radar experiment is presented to assess the availability and the performance of the developed radar sensor together with two distinct signal processing algorithms to detect and extract the main micro-Doppler features of human walk, namely torso and arms. The obtained results demonstrate the accurateness of the developed sensor and the success of the applied techniques with good fidelity.

Ensuring reliable satellite communication is critical for mission success, particularly for nanosatellite operations where power and resource constraints pose significant challenges. This paper validates the required ground station (GS) transmit power for the RIDU-Sat 1 TT&C uplink by combining theoretical analysis with dynamic simulation. A link budget was first calculated to determine the minimum required power, showing that a GS with an Effective Isotropic Radiated Power (EIRP) of 23.19 dBm is theoretically needed to meet the satellite's -126 dBm sensitivity threshold. To verify this and account for operational margins, a higher operational EIRP of 39 dBm was chosen for simulation. An end-to-end GNU Radio simulation modeling a full Low Earth Orbit (LEO) pass was then developed. The simulation results confirm that transmitting at 39 dBm successfully maintains the received signal strength at the satellite well above the -126 dBm threshold, thereby validating the link budget calculation and confirming the viability of the chosen operational power.

Indonesia's unique geography makes it vulnerable to natural disasters, compromising the terrestrial communication infrastructure. The Low Earth Orbit (LEO) satellite-based LoRa network offers a promising support solution for data communications, especially for transmitting critical disaster-related information such as seismic activity. However, managing uplink access from numerous sensors using conventional Medium Access Control (MAC) protocols faces challenges with collisions and efficiency. This paper proposes and evaluates Beacon-Synchronized TDMA (B-STDMA), a novel scheduling scheme designed to enhance uplink performance for Indonesian disaster communication via LEO satellite-LoRa. In B-STDMA, the satellite broadcasts a beacon for terminal synchronization, and pre-assigned unique terminal IDs determine dedicated, collision-free transmission slots. Using MATLAB-based simulations, the performance of B-STDMA is rigorously compared against Pure ALOHA and Slotted ALOHA, analyzing throughput, average packet latency, packet delivery ratio, and collision rates under varying network loads and device densities.

The effectiveness of satellite IoT for disaster monitoring is often limited by the data throughput bottleneck of traditional single-channel receivers, especially during mass reporting events. This paper addresses this bottleneck by presenting the design and verification of a low-cost, dual-channel LoRa receiver using Commercial-Off-The-Shelf (COTS) components. The primary contribution is the use of an ESP8266 microcontroller with a Software Serial implementation to manage two independent LoRa modules, overcoming the single hardware UART constraint of the MCU. System-level tests demonstrate that the hardware confirms accurate and isolated operation on two distinct channels (920 MHz and 923 MHz). It also showed that successful simultaneous data acquisition from both channels was achieved with consistent performance. This study validates the architecture could provide a low-cost, scalable solution that effectively doubles the data acquisition capacity, enhancing the robustness of LEO satellite missions for disaster response.

With the continuous rise in air traffic, the need for reliable and efficient assistive conflict detection systems has become highly crucial. This paper presents a deterministic predictive aircraft conflict detection approach that involves short-term trajectory prediction followed by the conflict detection process in two phases: a broad phase and a narrow phase. The broad phase uses swept Axis-Aligned Bounding Box volumes, applied for the first time in this field, generated around the predicted aircraft trajectories to identify potential conflicts. The narrow phase then performs step-by-step distance checks to confirm whether an actual conflict exists. Additionally, the system is designed to trigger alerts in the event of a loss of separation when a violation of separation standards occurs. The system was implemented in the BlueSky open-source air traffic simulator and evaluated using a variety of test cases reflecting possible aircraft conflict cases, generated within the simulator. The system demonstrated reliable conflict detection capabilities across most scenarios, offering a simple yet effective framework for en-route conflict detection and highlighting its potential for future development in tactical aircraft conflict detection research.

Shipboard landing of UAVs is a highly challenging problem due to the ship's motion, limited deck space, and environmental disturbances. This paper presents the development of a gain-scheduled Linear Quadratic Regulator (LQR) controller for a tilt-rotor UAV to perform autonomous shipboard landing. The UAV dynamics are modeled in the longitudinal axis, and ship motion is generated using a JONSWAP spectrum combined with Response Amplitude Operators (RAOs). Simulation results show that the proposed controller achieves accurate tracking of ship deck motion under Sea State 5 conditions.

Accurate mangrove classification using optical remote sensing remains challenging due to spectral similarity with other vegetation types, leading to significant spatial classification errors. This study evaluates four vegetation indices (NDVI, SAVI, CMRI, MVI) for mangrove-non-mangrove discrimination in Demak, Central Java, using Sentinel-2 imagery through comprehensive accuracy assessment incorporating histogram analysis, confusion matrices, and area-based spatial error evaluation. Results demonstrate SAVI's superiority with 79.3% overall accuracy and 79.9% producer's accuracy, attributed to its soil brightness correction mechanism addressing coastal spectral complications. NDVI achieved exceptional user's accuracy (97.3%) but moderate overall performance (63.9%), while CMRI showed poorest results (51.2%) due to complex trimodal distribution patterns. MVI exhibited conservative classification with minimal commission errors but significant omission issues. Spatial analysis revealed misclassification clustering at ecosystem boundaries, particularly mangrove-aquaculture transitions, with SAVI achieving highest true positive detection (106.1 hectares) and lowest omission errors (13.2 hectares). This study provides evidence-based guidance for automated mangrove monitoring systems and establishes a replicable evaluation framework for operational applications.

Viewing the Urban Heat Island (UHI) phenomenon through the Local Climate Zone (LCZ) framework offers a refined perspective for understanding urban thermal dynamics. This morphology-based approach enables systematic analysis of spatial and seasonal UHI variations. This study maps and evaluates seasonal UHI patterns in the Jakarta metropolitan area (Jakarta, Depok, Tangerang Raya, and Bekasi Raya) during 2014-2024. Land Surface Temperature (LST) was derived from Landsat 8 OLI/TIRS using the mono-window algorithm, while LCZ classification was obtained from the World Urban Database Portal and Access Tool (WUDAPT). Seasonal LST composites- December-February (DJF), March-May (MAM), June-August (JJA), and September-November (SON)-were averaged over 10 years to produce stable thermal patterns. UHI intensity was calculated as the LST difference between each pixel and the rural reference zone, then analyzed by LCZ class. Results show that LCZ 3 (compact low-rise) consistently recorded the highest LST, reaching 45.7 C in Jakarta during SON, followed by MAM as the second hottest season. The highest seasonal UHI occurred in LCZ 8 (large low-rise) in Bekasi Raya during MAM (8.64 C), with Jakarta (SON), Tangerang Raya (MAM), and Depok (MAM) in LCZ 3 peaking at 6.43 C,7.54 C, and 5.60 C, respectively. Dense built-up LCZs (LCZ 1-3, LCZ 8, LCZ 10) contributed most to UHI intensity, while vegetated zones (LCZ 11-14,16-17) consistently exhibited negative values. These findings confirm the LCZ approach's effectiveness in capturing thermal heterogeneity and supporting climate-sensitive urban planning in tropical megacities

Illegal waste disposal is a persistent issue in Indonesia. This study explores the use of medium-resolution satellite data-specifically Landsat 8, Sentinel-2, and Sentinel-1-for landfill detection in large sites such as Tempat Pembuangan Akhir (TPA). For optical data, the method employs surface reflectance values, a decision tree classifier, and Otsu thresholding, with NDVI applied to separate waste from vegetation. The Blue band proved particularly effective for distinguishing landfill from bare land. The approach achieved over 90% overall accuracy for both Landsat 8 and Sentinel-2, demonstrating strong potential for monitoring large landfill areas, though further testing is needed for smaller sites. Sentinel-1 data, while capable of capturing structural changes through multi-temporal analysis, showed limited ability to differentiate landfill from other solid objects when using single-date imagery.

Land subsidence is a growing threat to Greater Jakarta's infrastructure and public services, driven by excessive groundwater extraction, rapid urbanization, and soft alluvial geology. This study uses multi-year InSAR time-series data (2017-2023) to map vertical displacement and quantify exposure across transport networks, public facilities, and transport hubs. Results show severe and spatially extensive subsidence, with highways and primary roads most affected-several kilometres lie in zones subsiding faster than −20 mm/yr. Commuter rail and the Jakarta-Bandung High-Speed Rail also traverse moderate to high subsidence areas, while hospitals emerge as the most exposed public service. Although most airports are stable, localized high-subsidence pockets near Halim Perdanakusuma and Budiarto airports pose operational risks. The findings underscore the need to integrate subsidence monitoring into infrastructure planning, improve groundwater management, and prioritize mitigation in high-risk corridors.

Multifunction radar must operate reliably in multitarget environments, where waveform design must balance resolution, robustness, and interference suppression. Circulating code orthogonal frequency-division multiplexing multiple-input multiple-output (CC-OFDM MIMO) is a promising approach, but most prior studies have been restricted to single-target cases. This paper evaluates the performance of three coded waveform families-Golay, Zadoff-Chu (ZC), and M-sequence-in CCOFDM MIMO radar under multitarget conditions. A mathematical model is formulated to capture inter-target interference and thermal noise, and simulations with five targets are performed in closely spaced range-angle scenarios. Results demonstrate that Golay achieves superior sidelobe suppression, ZC provides robustness against frequency shifts with the highest SINR ceiling, and M-sequences offer acceptable statistical behavior but with greater leakage. These insights establish clear trade-offs among code families and provide practical guidelines for waveform selection in multifunction CC-OFDM MIMO radar.

Aviation safety remains a top priority, especially as global air traffic continues to increase. Traditional air traffic surveillance technologies, such as radar, are limited in both range and accuracy, particularly in remote regions and over oceanic areas. To address these limitations, Automatic Dependent Surveillance-Broadcast (ADS-B) has emerged as a more accurate, satellite-based surveillance solution. This study focuses on the design and development of an ADS-B system, emphasizing the implementation of a 4-way wilkinson power divider. The system is intended to connect four antenna arrays. The design process begins with defining specifications, calculating dimensions, and simulating a 2-way power divider using CST Studio 2019, followed by optimization to meet performance criteria. The 4-way configuration is derived without additional recalculations. A prototype is fabricated and tested using a Lite VNA, with measured results compared against simulation data. Evaluation is conducted to analyze discrepancies and assess system performance.

Indonesia ranks among the most disaster-prone countries globally, recording 1,162 disaster events in May 2025 alone, including 779 floods and 207 extreme weather incidents, resulting in 211 fatalities and displacing over 3.4 million people. These figures highlight the urgent need for a reliable communication system, especially when conventional networks fail during disasters. This study introduces BATAR (Bantuan Tanggap Radio), an integrated satellite-based emergency communication platform combining the Quadrifilar Helix (QFH) antenna, Automatic Packet Reporting System (APRS), and a voice repeater system. The BATAR antenna is designed with an omnidirectional radiation pattern and performs optimally in the VHF and UHF bands, enabling stable satellite connectivity without the need for a rotor-based tracking system. Compared to traditional Yagi or aluminum-pipe QFH antennas, BATAR offers greater portability, ease of assembly, and resilience to extreme weather. APRS supports real-time data transmission-such as location, text, and weather-critical for search and rescue and disaster response. The LAPAN-A2 (IO-86) satellite, which passes over Indonesia up to 14 times daily, provides consistent voice communication support when ground infrastructure is unavailable. Moreover, the BATAR antenna is compatible with the RIDU-Sat 1 satellite, expanding its potential for satellite-based emergency communication. Supported by portable and energy-efficient hardware, BATAR ensures quick deployment and real-time monitoring in the field. The integration of these technologies within the BATAR platform offers a scalable and resilient solution for disaster communication, aiming to enhance coordination and response during emergencies across vulnerable regions of Indonesia.

Indonesia, as the world's largest nickel producer with 1.8 million tons of output and reserves exceeding 55 million tons, has implemented downstream processing policies to enhance resource value. This study analyzes the spatial distribution of nickel mines, smelters, and supporting infrastructure in Sulawesi and Maluku Islands using geospatial visualization and cluster analysis. Data sourced from the Ministry of Energy and Mineral Resources, Geospatial Information Agency, Ministry of Transportation, and Marine Traffic were integrated and processed according to the Knowledge Discovery in Databases framework. Thematic maps in QGIS reveal concentrated mining and smelting activities in Morowali and Konawe, Sulawesi, leveraging 44 active ports and extensive road networks, contrasting with dispersed operations and infrastructure deficits in Maluku. A Density-Based Spatial Clustering of Applications with Noise (DBSCAN) applied to normalized distance measures indicates that 98 percent of mining sites are proximate to smelters, and 96 percent of smelters have adequate road and port access. Outlier clusters highlight areas where additional facilities are needed. Despite substantial foreign investment-predominantly Chinese-environmental and social challenges such as illegal deforestation, marine pollution, and reliance on coal power require attention. These insights inform targeted infrastructure development and policy interventions to promote sustainable and equitable downstreaming of Indonesia's nickel industry, inclusive economic growth.

Artificial intelligence (AI) has recently become an essential tool in the development of drone detection algorithms, particularly for classifying passive RF signals that are often affected by background noise or interfering sources such as Wi-Fi transmissions. This study focuses on an optimization of sliding window overlaps on deep learning-based drone signal classification. Complex-valued I/Q data were collected in real-world environments using low-cost SDR hardware and segmented into fixed-length frames with varying overlap ratios. A convolutional neural network, developed by MATLAB, was trained to each configuration using the trainnet function. It was observed that a 25% overlap provides the best trade-off between data efficiency, defined as effective learning from limited data, and classification performance.

This study explores the fusion of RGB and Near-Infrared (NIR) images for scene classification. Three fusion techniques, such as RedSwap, Heatmap, and the proposed NIRR Difference, were evaluated using visual quality metrics. A 4-channel image set was first classified using an Intermediate Fusion conventional neural network (CNN). Due to overfitting, a Patch-Based Intermediate Fusion CNN was introduced, where fixed-size patches were processed to improve generalization. The patch-based model achieved higher validation accuracy (up to %61) and reduced overfitting. Experimental results, visual outputs, and metric-based evaluations confirm the effectiveness of the proposed approach in image fusion and classification.

Accurate battery capacity prediction is crucial for ensuring the performance and mission reliability of satellites like LAPAN-IPB/LAPAN-A3, facilitating efficient energy management and enabling early detection of potential system failures. Operating in a Low Earth Orbit (LEO) with a polar orbit, the LAPAN-IPB satellite's Lithium-ion battery undergoes regular charge and discharge cycles, influenced by its orbital path relative to the Sun and Earth. This study introduces a novel hybrid deep learning approach for predicting the LAPAN-IPB satellite's battery capacity, leveraging historical telemetry data. Drawing inspiration from successful applications in Short-Term Load Forecasting (STLF), where hybrid SARIMAX-Deep Learning methods have excelled in handling time series data with seasonal and non-linear patterns, this research combines the strengths of traditional statistical modeling with advanced deep learning techniques. Specifically, the SARIMAX model is first employed to capture linear patterns and seasonal trends within the battery capacity data. The residuals (prediction errors) from this SARIMAX model, which contain complex non-linear information that SARIMAX alone cannot fully capture, are then fed into deep learning algorithms-including Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and Bi-directional LSTM (BiLSTM). Among the evaluated hybrid models, the SARIMAX-LSTM model demonstrated superior performance, achieving the highest R2 score of 0.92131 and a Mean Squared Error (MSE) of 0.013787, affirming its effectiveness for precise satellite battery capacity prediction

Climate change has emerged as a critical global challenge, with far-reaching consequences for our planet's ecosystems, weather patterns, and human societies. With the advancement in technology today, decision-makers have begun to move away from costly and traditional monitoring methods to leveraging more advanced techniques such as remote sensing for agricultural monitoring. Traditional methods of assessing agricultural lands often involve time-consuming and costly field surveys. However, with recent advancements in remote sensing technology and with the availability of satellite imagery, it has become possible to utilize remote sensing indices that assess agricultural lands more efficiently and accurately. This study focuses on utilizing Sentinel-2 and Landsat-8 satellite images to provide real-time and multi-dimensional insights on a seasonal basis. The primary objective of this study is to assess the agricultural lands in the Northern Governorate of the Kingdom of Bahrain by analysing the potential of remote sensing indices in characterizing and quantifying key agricultural parameters such as vegetation health, water stress, temperature levels and soil conditions. To extract meaningful geospatial insights from satellite imagery, various remote sensing indices have been applied such as the Normalized Difference Chlorophyll Index (NDCI), Normalized Difference Moisture Index (NDMI), Soil Moisture Index (SMI), Soil Salinity Index (SSI) and Land Surface Temperature (LST). The comparative analysis reveals key spatial and temporal dynamics between all indices. Correlation coefficient (r) highlights relationships between NDCI, SMI, and NDMI, showing notable interdependence in agricultural parameters with r values ranging from 0.623 to 0.783 across summer and winter seasons.

Tripe Jaya Sub-district, located in Gayo Lues Regency, has the characteristics of a hilly area with high rainfall, steep slopes, and often experiences landslides and erosion especially along the road. The purpose of the study was to determine the relationship between landslides and erosion in variables of each class of landslide vulnerability scoring method with erosion hazard class USLE method in bivariate correlation. The results showed that landslide prone areas in Tripe Jaya Sub-district are divided into five classes, namely very low (279.21 Ha), low (3,639.80 Ha), medium (17,893.07 Ha), high (19,822.91 Ha), and very high (150.23 Ha). Meanwhile, the erosion hazard level is also divided into five classes: very light (2,909.09 Ha), light (20,669.38 Ha), medium (10,880.66 Ha), heavy (432.99 Ha), and very heavy (6,922.76 Ha). The correlation test results show that the very low landslide class and the very light erosion class have a correlation value of 0.07, indicating little if any relationship. However, correlations between other landslide classes (low to very high) and erosion classes (High to very High) showed values of 0.00-0.01, with relationship coefficients ranging from 0.73-0.98, indicating strong to very strong relationships.

Floods are the primary hydrometeorological disaster in Indonesia, with rainfall serving as a key controlling factor. Beyond environmental damage, floods cause significant human casualties across affected regions. Satellite data can effectively map flood inundation patterns. The SATGPT platform, based on Joint Research Center (JRC) Global Surface Water Mapping data, quantifies how frequently areas experience flood inundation over specified time periods, excluding permanent water bodies such as rivers and lakes. This frequency map serves as a flood hazard map indicating where floods commonly occur. This study validates the SATGPT flood inundation map against historical flood records from the National Disaster Management Agency (BNPB) and determines rainfall thresholds from Climate Hazards Group InfraRed Precipitation with Station (CHIRPS) data that trigger flood events. The validation methodology employs spatial buffering of flood geo-location data combined with descriptive and predictive statistical analyses, including mean, median, standard deviation, and percentile calculations. Results demonstrate that the SATGPT flood map achieves 83% accuracy when compared to BNPB historical data from 2021-2025, particularly for January when most floods occur. The SATGPT data successfully represents historical flood events across various Indonesian regions. Using percentile analysis, the threshold for extreme rainfall associated with flood conditions is determined to be approximately 46 mm/day based on CHIRPS data. However, this calculation covers the entire Indonesian territory without accounting for regional climate variability. These findings demonstrate the potential of remote sensing technology for flood mapping and early warning systems in flood-prone communities.

The increase in the number of motor vehicles is not proportional to the availability of transportation facilities and infrastructure, which can increase the risk of traffic accidents in the city of Yogyakarta. Therefore, determining accident-prone locations can be one of the efforts to reduce the number of accidents in the city of Yogyakarta. This study aim to map the distribution of traffic accident risk levels on certain sections of roads in Yogyakarta City using a weighted hierarchical analysis method. The results of the study indicate that there are four road segments that are in a highly accident-prone condition, namely HOS Cokroaminoto road of segments 1 and 2, Pangeran Diponegoro road of segment 3, and Menteri Supeno road.

In this paper, a compactly designed antipodal Vivaldi antenna (AVA) surrounded by a 3D-printed PLA cover is proposed to make the antenna compatible with Ground Penetrating Radar (GPR) applications in the frequency interval of 1.54.5 GHz. Compared to the bare antenna, the lower frequency is decreased by 200 MHz, and the gain is increased by 1 dB, utilizing full-wave simulations and optimizations. In these optimizations, PLA boxes of various thicknesses and fill rates are investigated. Furthermore, a reduction in side lobe levels is observed. Based on the results obtained, the antenna was deemed suitable for use in GPR applications.

Mapping aboveground carbon stock (AGC) in seagrass beds through remote sensing is challenging and requires integration with field data. Acquiring field data for seagrass AGC is resource-intensive, time-consuming, expensive, and potentially damaging to the environment. Consequently, there is a need to develop an approach that enables non-destructive mapping of seagrass AGC, encompassing both field surveys and image processing. The objective of this study is to establish an equation that can predict seagrass AGC based on their percent cover (PC) for various seagrass species, including Enhalus acoroides (Ea), Thalassia hemprichii (Th), Thalassodendron ciliatum (Tc), Cymodocea rotundata (Cr), Halophila ovalis (Ho), and Syringodium isoetifolium (Si). The derived species-specific PC to AGC equations are then applied to transform field PC values into AGC. These modelled AGC values are utilized to train a PlanetScope 8-bands SuperDove imagery using a stepwise regression model in the Nemberala region of Rote Island. Our findings reveal a strong relationship between seagrass PC and laboratory-measured AGC, with R-squared values higher than 0.5 for all six species, with the highest at 0.84 for Cr. The regression function for estimating seagrass AGC from the PlanetScope imagery incorporates four spectral bands: Green I, Coastal blue, Yellow, and NIR, with a correlation coefficient of 0.59 and an R-squared value of 0.34. Based on these estimations, the Root Mean Square Error (RMSE) value for seagrass AGC mapping is calculated as 1.03 grams of carbon per square meter, and the total seagrass AGC in the study area is projected to be 0.97 tons.

Monitoring corn crop phenology is vital for agricultural management but often hindered by labor-intensive data collection, especially over large areas. Remote sensing, particularly Synthetic Aperture Radar (SAR), offers an efficient alternative. This study utilized Sentinel-1 SAR imagery from October 5, 2023, to February 2, 2024, with 12-day acquisition intervals. After preprocessing and processing, corn phenology was characterized using two smoothing techniques: Locally Estimated Scatterplot Smoothing (LOESS) and the Savitzky-Golay Temporal Filter. Results showed varying peaks and troughs in phenology trends, with R² values differing across parameters. The Savitzky-Golay filter performed better, yielding higher R² values. Temporal markers from local extrema identified key phenological stages. Among parameters, Alpha exhibited the highest R² (0.96) and lowest RMSE (0.02), along with a smaller interquartile range, indicating stable performance over time. Phenological stage transitions were determined through breakpoints, inflection points, quarter dates, mid dates, and local extrema. Quarter dates divided growth stages from vegetative to reproductive phases: the first quarter aligned with the Tenth Leaf stage, the second with Blister, the mid date between Blister and Dough, and the final quarters near Dent. Overall, Sentinel-1 SAR proved effective in monitoring corn crop phenology, with Alpha emerging as the most reliable predictive parameter. This approach demonstrates the potential of SAR-based monitoring to reduce manual fieldwork and improve the precision of crop growth assessments.

Ground-based synthetic aperture radar (GB-SAR) systems are primarily used to enhance the understanding of the complex mechanisms underlying microwave backscattering. In this study, a polarimetric calibration procedure is applied to GB-SAR data, effectively reducing channel imbalance during post-processing and thereby improving image quality. Two types of corner reflectors (CRs)-a dihedral reflector oriented at 0 degree and 45 degree and a trihedral reflector-are employed as reference calibration targets for polarimetric radar calibration methods. Subsequently, small unmanned aerial vehicle (UAV) test objects are analyzed. The results demonstrate that the polarimetric signatures obtained after calibration accurately correspond to the expected scattering responses of the test target.