Journal of Military Science and Technology

Effect of inorganic filler type and in-situ Silane modification on lap shear strength and thermal aging durability of polyamide-cured epoxy adhesive on aluminum alloy

2026-07-01T01:56:23+00:00

This study investigates the effects of inorganic filler type (ZrO₂, Al, TiO₂, TiC, WC, h-BN) and in-situ silane modification (APTMS, GPTMS, VTMS) on the lap shear strength (LSS) and thermal stability of a polyamide-cured epoxy adhesive on 6061 aluminum alloy. The optimal ZrO₂ content was 5.0 wt%, yielding an initial LSS₀ of 12.4 MPa and a 500-h thermal retention factor at 80 °C (TRF-500h) of 83.1% without silane. APTMS outperformed other silanes through a dual interfacial bonding mechanism, raising the LSS₀ to 15.6 MPa and TRF-500h to 88.5% for the ZrO₂-5.0-APTMS system. Evaluation of six filler types against dual criteria (LSS₀ ≥ 15 MPa, TRF-500h ≥ 85%) identified h-BN, WC, ZrO₂, and TiC as qualifying systems. The h-BN-5.0-APTMS system achieved the best overall performance (TRF = 90.6%, LSS₅₀₀h = 14.5 MPa), driven by the unique tortuosity effect and high thermal conductivity of its lamellar structure. TGA/DTG analysis confirmed higher T₅% values for all filled systems compared to the unfilled epoxy (323 °C), following the order WC (341 °C) > BN (339 °C) > TiC (337 °C) > ZrO₂ (331 °C) > Al (329 °C) > TiO₂ (325 °C) due to the oxygen diffusion barrier effect. These results guide the development of thermally durable modified adhesives for structural applications.

Research on the utilization of fly ash and rice husk silica as fillers for flame-retardant silicone rubber materials

2026-07-01T01:56:56+00:00

This paper investigates the effects of halogenated flame retardants (DBDPO, DBDPE, and Chloroparaffin), halogen-free fillers derived from industrial and agricultural waste (Fly Ash and Biosilico amorphous silica from rice husk), and the anti-structuring silanol agent ND-8 on the fire resistance of silicone rubber materials. The results demonstrate that incorporating 100 phr (parts per hundred rubber) of Fly Ash combined with Biosilico amorphous silica and silanol ND-8 into the silicone rubber compound facilitates the creation of a novel material with superior fire resistance. This improvement is attributed to the formation of a stable mullite and cristobalite ceramic surface layer, resulting in a residual ash content of 60% upon combustion. The material achieved a 5 VA rating (the highest level under the UL-94 standard), while simultaneously contributing to resource conservation and environmental pollution mitigation.

Advanced ZnO:Ce@GO nanocomposites for enhanced visible–light photocatalytic degradation of Disperse Red 152

2026-07-01T01:59:38+00:00

ZnO and ZnO:Ce@(0–8)%GO nanocomposites were synthesized hydrothermally and characterized by XRD, FT–IR, PL, EDS, and UV–Vis–DRS. The materials retained the wurtzite ZnO structure with crystallite sizes of 28–38 nm. Ce doping and GO incorporation reduced the band gap from 3.15 to 2.78 eV and significantly suppressed electron–hole recombination, as evidenced by PL quenching. Under visible-light irradiation, ZnO:Ce@8%GO achieved 98.60% degradation of DR152 after 210 min, compared with 79.69% for pristine ZnO, with a reaction rate 2.94 times higher. The composite also retained 83.71% of its activity after five cycles, demonstrating good stability and the synergistic enhancement of photocatalytic performance by Ce and GO.

The incorporation of water-soluble Hydroxypropyl-β-Cyclodextrin-Lycopene inclusion complex in biopolymer chitosan nanoparticles: Physicochemical characteristics and in-vitro release

2026-07-01T01:59:09+00:00

Lycopene (Lyc) from Gac fruit is a potent antioxidant, but its poor water solubility limits its applications. To address this, Lyc was complexed with Hydroxypropyl-β-cyclodextrin (HPCD) and encapsulated in chitosan nanoparticles via ionic gelation. The optimal 1:1 Lyc-HPCD mole ratio achieved 63.0 ± 3.1% encapsulation, while chitosan-loaded complexes (127.6 nm, +21.38 mV) reached 68.79 ± 3.75%. Characterization by UV-Vis, FTIR, DSC, and SEM confirmed the structure, and the system showed sustained release over 24 hours at varying pH. This nanocarrier offers an efficient platform for lycopene delivery in food and cosmetic applications.

Fabrication of polypyrrole/nylon composite with high radar transmission loss for electromagnetic interference shielding application

2026-07-01T01:59:24+00:00

This study investigates the fabrication of conductive polypyrrole (PPy)/nylon composites via a controlled interfacial polymerization technique. The composite was synthesized by positioning a nylon fabric at the interface of an aqueous phase containing oxidant and p-toluene sulfonic acid (p-TSA, 0.1 M) as the dopant, and an organic phase consisting of 3% (v/v) pyrrole in hexane. The influence of oxidant types and reaction temperatures on the formation of the PPy layer was systematically evaluated. Morphological and structural characterization revealed that the PPy was successfully integrated into the nylon matrix, increasing the fabric thickness from 0.09 mm (raw) to 0.1 mm. Secondary electrochemical deposition was performed on the PPy/nylon composite in an aqueous solution containing 0.1 M pyrrole to achieve a greater PPy thickness. Furthermore, electromagnetic interference (EMI) shielding measurements in the X-band frequency range (8–12 GHz) showed a shielding effectiveness (SE) exceeding 20 dB. These results suggest that the as-prepared PPy/nylon composites are promising candidates for lightweight, flexible radar-absorbing materials in advanced electronic applications.

Research, design, and fabrication of a gas-generating propellant charge for a three-chamber underwater vehicle simulator launch system

2026-07-01T01:57:39+00:00

This paper presents selected results on the research, design, and fabrication of a gas-generating propellant charge using domestic technology, based on an internal ballistic model of a three-chamber launcher. Following the study and calculation of the igniter mass, propellant grains and the gas-generating charge were manufactured, with an igniter mass of 32,8-33,2 g and a total propellant mass ranging from 3580 g to 3590 g. Ground firing tests on a test stand met the required performance criteria. The charge was subsequently subjected to firing tests to evaluate its capability to propel an underwater vehicle simulator with a mass of 2200-2300 kg within a launch tube system. The results show that the pressure–time profiles of the propellant gases in the charge chamber (chamber 1), the accumulator (chamber 2), and the launch tube (chamber 3) are in complete agreement with theoretical predictions. The obtained internal ballistic performance parameters include the maximum propellant gas pressures in the charge chamber, the accumulator, and the launch tube, which are 180,21 bar, 92,56 bar, and 11,04 bar, respectively. The operating time of the propellant charge is 390 ms, and the exit velocity of the vehicle simulator from the launch tube reaches 21,95 m/s, satisfying the technical requirement.

Unsupervised deep learning-based anomaly detection in mixed gases using sensor array

2026-07-01T01:55:17+00:00

Anomaly detection in mixed-gas data is a significant challenge for multisensor systems because of their nonlinear characteristics, noise, environmental influences, and scarcity of anomalous data. This study proposes an unsupervised deep-learning anomaly detection method. An unsupervised Autoencoder model was employed to reconstruct two-dimensional image representations of normal mixed-gas data, thereby eliminating the need for anomalous labels during training. To simulate realistic abnormal conditions, anomalous samples were generated using diverse and physically meaningful perturbation transformation methods. An adaptive threshold based on the reconstruction error was used as the criterion for distinguishing between normal and anomalous data points. The effectiveness of the proposed method was evaluated using accuracy and F1-score metrics, which achieved values of 0.97 and 0.92, respectively. The experimental results demonstrate that the proposed approach can effectively learn the characteristics of normal data and accurately detect anomalies in mixed gas datasets. This method improves the safety, reliability, and intelligent monitoring capabilities of gas-sensor array systems in practical applications.

Image fusion in a man-portable multi-sensor optoelectronic device based on an embedded system on module

2026-04-14T04:24:49+00:00

Man-portable optoelectronic devices combining visible (or/and low-light) and infrared (IR) sensors require real-time image fusion to enhance situational awareness under degraded visual conditions, while operating under strict size, weight, power consumption, and cost (SWaP-C) constraints. This paper presents the design and implementation of a real-time EO/IR image fusion system based on an embedded System-on-Module (SoM) architecture for a handheld multi-sensor platform. The proposed system supports synchronized acquisition from multiple visible and IR cameras and executes parallel fusion processing using a multi-core CPU with GPU/NPU acceleration. Experimental results demonstrate real-time performance with end-to-end processing latency below 40 ms, sustained frame rates of 25 fps for fused video output. The fused image exhibits better contrast, clearer details, and improved target differentiation compared to the original input images. The experiments indicate that SoM-based architectures can provide a balanced and practical solution for real-time multi-channel EO/IR image fusion in man-portable applications.

Temperature-dependent XAFS parameters of crystalline platinum within a classical anharmonic correlated Einstein approach

2026-07-01T01:56:42+00:00

Temperature-dependent thermodynamic XAFS parameters of crystalline platinum (Pt) are investigated using a classical anharmonic correlated-Einstein approach that accounts for thermal disorder. The method is based on an anharmonic effective potential and classical statistical theory, from which effective force constants, Einstein temperature, and Einstein frequency are consistently obtained. Anharmonic lattice vibrations are treated using a fourth-order cumulant expansion, yielding analytical expressions for the first four XAFS cumulants over the temperature range 0–800 K. The results show good agreement with experimental data and quantum anharmonic-correlated Einstein calculations in the intermediate- and high-temperature regimes. Higher-order cumulants play a dominant role in describing anharmonicity and non-Gaussian features of the atomic distribution, remaining valid over a wider temperature range, including part of the region below the Einstein temperature. The present approach therefore provides a consistent and efficient description of temperature-dependent thermodynamic XAFS parameters of Pt.

Integrated guidance and control for UAV path following in Frenet–Serret frame

2026-07-01T02:00:26+00:00

This paper proposes a high-precision path following method for unmanned aerial vehicles (UAVs) using the Frenet–Serret (FS) coordinate frame combined with the optimal Integrated Guidance and Control (IGC) algorithm. To increase adaptability and reduce tracking error in an uncertain environment, a neural network coupled model predictive controller (NNMPC) is directly integrated into the IGC framework. The stability analysis of the system is performed using the Lyapunov–Krasovskii criterion, providing a sufficient condition for asymptotic stability under bounded time delay and neural network approximation errors. The simulation results for a quadrotor UAV following a sinusoidal trajectory demonstrate that the proposed method outperforms traditional PID control in terms of accuracy and disturbance rejection capability. Furthermore, this study discusses practical implementation constraints, including computational complexity and data availability, thereby bridging the gap between theoretical design and real-time UAV applications.

Measuring the global and directed propagation of several quality indicators in the optimal design of parallel robots through the evaluation of Lipschitz and GTI indices

2026-07-01T01:59:52+00:00

In the optimal design of multi-objective parallel robots, transmission ratios such as n_vmin and n_pmin are used to evaluate the extent to which velocity and force propagate from the joint space (q) to the robot's work space (x). Two pieces of information need to be measured in this evaluation: the degree of overlap between the drive and the actuator, and the amplification/attenuation of the drive when it becomes the actuator. Two methods are used to evaluate these factors: assessing the propagation across the entire space using the Lipschitz coefficient and evaluating in a specific direction of interest using the GTI (Geometric Transmission Index). In the improved Atlas algorithm proposed by the author, the Pareto boundary is defined as the intersection of convergent directional cones based on the propagation guidance atlases of the aforementioned design parameters. This paper introduces a method for evaluating these coefficients in the discrete design domain ri to automate the process of determining the Pareto boundary, thereby providing a basis for automating the entire design.

An improved sliding mode control combined with backstepping techniques and artificial neural networks for a coupled-tank system

2026-07-01T02:00:41+00:00

This study proposes a solution to design a liquid level tracking controller for a coupled-tank system (C-TS) using a sliding mode control (SMC) method based on proportional integral (PI) sliding surface (SS) combined with backstepping techniques and radial basis function neural networks (RBFNNs). The SMC controller based on proportional integral sliding surface (also called PISMC) provides more parameters with which to tune the SMC controller. The backstepping approach ensures the global asymptotic stability of strict feedback systems. The improved backstepping sliding mode control is employed to enhance the performance of conventional sliding mode control and act as a robust control strategy. The RBFNNs are utilized to approximate the unknown nonlinear functions in the improved backstepping sliding mode control. System stability is proven through Lyapunov theory. Simulation results in MATLAB/Simulink demonstrate the effectiveness, appropriateness, and robustness of the proposed control method in comparison with the fuzzy controller, adaptive control using RBF neural network, sliding mode control (SMC) based on disturbance observer and Quasi mode and sliding mode control using conditional integrators with the rising time achieves 0.1295 (s), the settling time is 0.2233 (s), the percent overshoot is 0 (%), the steady state error converges to zero.

Improving the performance of convolutional neural networks using a fuzzy logic based pooling method

2026-07-01T01:58:53+00:00

In automatic image recognition and classification tasks, convolutional neural networks (CNN) are widely used, in which pooling operations play an important role in aggregating and highlighting image feature characteristics. This paper proposes a fuzzy logic based pooling method that integrates minimum, average, and maximum feature responses through a simple Takagi–Sugeno inference mechanism. Unlike conventional pooling methods, the proposed approach explicitly models uncertainty in feature aggregation, enabling a better balance between salient local features and contextual in order to preserve important image information. The effectiveness of the proposed method is evaluated on MNIST and CIFAR-10 datasets. Experimental results demonstrate that the proposed fuzzy pooling method achieves superior classification performance compared to conventional pooling methods, while maintaining comparable computational complexity. This enables the method to be effectively applied in CNN based image classification tasks.

Integrated Probabilistic Engagement Modeling and Guidance Accuracy Optimization for Launch and Lethal Zone Construction

2026-07-01T01:58:24+00:00

This paper develops a unified probabilistic framework for missile–target engagement assessment integrating kinematic reachability, fragment-based lethality modeling, and stochastic terminal guidance uncertainty. Closed-form approximations for expected kill probability are derived under Rayleigh-distributed miss distance. Analytical asymptotic analysis reveals an exponential dependence of engagement effectiveness on inverse-squared guidance accuracy. A deterministic lethal-radius model is compared with the proposed probabilistic formulation, demonstrating significant overestimation by traditional methods. An optimization problem is formulated to determine the minimum required terminal guidance accuracy ensuring a prescribed kill probability threshold. The results provide quantitative design constraints for guidance and control system development.

Proposed early-stopping algorithm for the FPGA implementation of a Min-Sum decoder for NB-LDPC codes

2026-07-01T01:57:53+00:00

This paper proposes an early-stopping algorithm applied to the design of a Min-Sum decoder for NB-LDPC (32,16) codes on FPGA. The decoder is implemented in SystemVerilog for NB-LDPC codes over GF(16) and evaluated on AWGN channels with performance metrics including bit error rate (BER), frame error rate (FER), average iteration count, and decoding latency. The early-stopping algorithm is integrated into the loop control block to terminate the decoding process as soon as convergence is achieved, thus reducing the average number of decoding iterations. Implementation results show that the average decoding latency is reduced to 134.4 ns, corresponding to BER ≈ 5.31 × 10⁻³and FER = 0.12, consistent with reference MATLAB simulations. These results demonstrate that implementing NB-LDPC coding and decoding algorithms on FPGA, combined with an appropriate early-stopping mechanism, can significantly reduce hardware processing time, enabling real-time communication applications.

Robust and lightweight UAV visual localization in GNSS-denied environments using a variational autoencoder

2026-07-01T01:58:08+00:00

This paper proposes a robust and lightweight visual localization framework for UAVs in GNSS-denied environments. Utilizing a Variational Autoencoder (VAE) trained on full RGB imagery, the system extracts rich features compressed into an optimized 256-dimensional latent space to accommodate onboard constraints. These features are matched using an unnormalized Euclidean distance, while a Linear Kalman Filter (LKF) smooths the trajectory. Experiments demonstrate this model outperforms baselines, achieving a raw RMSE of 0.087 m, which improves to 0.065 m with the LKF. This approach ensures stable, highly accurate real-time navigation.

Using a MILP-based model for solving the optimal power flow problem in distribution networks integrated with step-voltage regulators

2026-07-01T01:57:10+00:00

This paper proposes a mixed-integer linear programming (MILP) model for the optimal power flow (OPF) problem in distribution grids with step-voltage regulators (SVRs). The OPF problem aims to minimize the total cost of the distribution grid, including the cost of purchasing effective and reactive power from the transmission grid, the cost of generating effective power, and the reactive power of distributed power sources. The constraints considered include the power flow equations, the node voltage constraints, the branch transmission power limits, the power factor limits at the connection point, and the SVR constraint. The MILP model of the OPF problem in the distribution grid was developed from the mixed-integer nonlinear programming (MINLP) model by linearizing the power flow equations and building an accurate linear SVR model. The proposed MILP model is evaluated on an IEEE33-node grid with different load scenarios using the GAMS programming language and the CPLEX commercial solver. The calculation results show that the MILP model is computationally efficient, and the optimal pressure distribution of SVR reduces the operating costs of the distribution grid.

Effective dielectric modeling for antenna lenses using fill-factor analysis and the Nicolson–Ross–Weir method

2026-07-01T01:58:38+00:00

The dielectric layer thickness is a key parameter in lens antenna systems, determining the controllability of phase distribution and radiation efficiency. However, the use of multiple materials with different dielectric constants often increases complexity and fabrication costs. To address this issue, this paper proposes establishing a streamlined and highly reliable design workflow for single-material lens antennas by rigorously validating a simple closed-form analytical method, based on the micro-periodic fill factor, against the complex Nicolson–Ross–Weir (NRW) parameter extraction algorithm. Computational and simulation results demonstrate that the effective dielectric values obtained from the analytical fill-factor approach exhibit excellent agreement with the NRW method, thereby proving that the proposed analytical model can effectively replace time-consuming full-wave parameter extraction processes and enable rapid optimization of the dielectric layer thickness. The method was applied to the design of a planar Lüneburg lens operating at 30 GHz, achieving an aperture efficiency above 0.9 across a broad frequency range while maintaining a stable beam pattern over various scanning angles. These results demonstrate the feasibility of the approach and highlight its potential for developing thin, easy-to-fabricate planar lens antennas suitable for high-frequency communication and radar systems.

Efficient solutions for reducing electromagnetic interference in legacy electronic test equipment systems

2026-07-01T02:00:09+00:00

This paper presents a comprehensive methodology for reducing electromagnetic interference (EMI) and improving electromagnetic compatibility (EMC) during the modernization of legacy analog-digital mixed electronic test systems. The research addresses critical challenges encountered when replacing obsolete Soviet-era discrete components with modern commercial off-the-shelf (COTS) integrated circuits in precision optoelectronic testing equipment. A systematic approach combining signal integrity analysis, power distribution network (PDN) optimization, and multi-layer shielding techniques is proposed. Experimental results demonstrate successful EMI suppression achieving conducted emission levels below CISPR 11 Class B limits (66 dBμV at 150 kHz to 40 dBμV at 30 MHz) and radiated emission compliance with MIL-STD-461G RE102 requirements. The proposed methodology enables functional equivalence while maintaining measurement accuracy within ± 0.5% of original specifications across the operational bandwidth of 1 MHz to 100 MHz.

Optimization of process parameters for hydromechancal deep drawing of square conical parts based on Taguchi and ANOVA methods

2026-07-01T02:00:55+00:00

In this study, the effects of process parameters on the forming height of square conical parts in hydromechanical deep drawing were investigated using the Taguchi experimental design combined with analysis of variance (ANOVA). SUS 304 stainless steel was selected as the blank material, and experiments were carried out on a hydromechanical deep drawing system equipped with square conical punch and die sets. Three main process parameters were considered, including the relative blank thickness (s/D), blank holding force (q_BHF), and initial forming fluid pressure (P₀). A Taguchi L9 orthogonal array was employed to evaluate the influence of these parameters on the product height based on the “larger-the-better” quality characteristic. The results indicate that the blank holding force has the most significant effect on the forming height, followed by the initial fluid pressure and the relative blank thickness. The optimal combination of process parameters was determined as s/D = 0.0067, q_BHF = 6 MPa, and P₀ = 25 MPa. The confirmation experiment under the optimal condition achieved a forming height of 45 mm, with a deviation of 2.97% from the predicted value, demonstrating good agreement between prediction and experiment. The findings of this study provide useful guidance for process parameter optimization in the hydromechanical deep drawing of square conical components

Statistical evaluation and regression modeling of strength recovery in scarf-repaired composite laminates

2026-07-01T01:55:39+00:00

Scarf repair is one of the most effective bonded repair techniques for damaged composite laminates because it can achieve high strength recovery while minimizing stress concentration. In this study, the influence of scarf geometry on the tensile behavior and strength recovery rate (SRR) of scarf-repaired composite laminates was investigated using a statistically supported approach. Experimental tensile-test data corresponding to scarf ratios of 1/10, 1/20, and 1/30, together with defect sizes of 6 mm and 12 mm, were analyzed using two-way analysis of variance (ANOVA) to evaluate the significance of geometric parameters on SRR. The results showed that scarf ratio is the dominant factor affecting repair performance (p<0.001), whereas defect size and interaction effects are statistically insignificant within the investigated range. The SRR increased significantly from approximately 51.0–52.9% at a scarf ratio of 1/10 to approximately 79.2–82.0% at scarf ratios of 1/20 and 1/30. An exponential decay relationship between SRR and scarf angle was established using nonlinear regression analysis (R²=0.898). Additional experimental results at a scarf ratio of 1/5 exhibited a noticeable deviation from the model prediction, indicating a transition in the dominant failure mechanism at relatively large scarf angles. The findings provide a statistically supported framework for evaluating scarf repair configurations and balancing mechanical efficiency with repair manufacturability.

Study on the effects of training round caliber and propellant type on the ballistic characteristics of a dynamic-reaction training device

2026-07-01T01:56:00+00:00

This study proposes the structural and ballistic design for an 82 mm PUI-type training device based on the dynamic-reaction principle. Using a MATLAB-based program to ensure the training projectile's trajectory matches that of the combat projectile, the authors analyzed the influence of two main parameters: the caliber of the training projectile (d) and the propellant type. The results show that NBL-62 propellant significantly reduces maximum pressure and stabilizes ballistic characteristics compared to NBL-34. The study identifies d = 23 mm with NBL-62 as the optimal configuration, ensuring a compact, low-pressure, and stable device. These findings provide an important scientific basis for designing and manufacturing an 82 mm training device based on the dynamic-reaction principle

A frame-level video annotation tool for dynamic gestures and apply in Vietnamese sign language

2026-07-01T01:57:24+00:00

Human Action Recognition (HAR) in video is essential for human–computer interaction, particularly in sign language and smart device control. However, model performance depends heavily on accurate temporal annotation of dynamic gestures. This study proposes a frame-level video annotation tool for Vietnamese sign language, enabling precise temporal segmentation and structured JSON export for deep learning applications. A dataset of 15 dynamic gesture classes is also constructed using a multi-view acquisition setup. Annotation quality is evaluated using Temporal IoU, Cohen’s Kappa, and annotation time. Results show high inter-annotator agreement at 0.9470 and 0.9888, respectively, demonstrating the effectiveness of the proposed tool for reliable and efficient gesture annotation.