Accurate estimation of aircraft aerodynamic and stability parameters plays a very important role in flight dynamics modeling, control system design and performance evaluation, especially in defence aviation and unmanned aerial vehicles (UAVs). Parameters such as lift, drag, moment coefficients, and control derivatives are typically derived from flight tests, but conducting such experiments can be expensive, risky, or restricted in defence settings. Existing methods, including gradient-based recurrent neural networks (RNN-G), face limitations in handling fast-changing, nonlinear dynamics owing to their static memory structures. This study proposes an Adaptive Forgetting Factor Recurrent Neural Network (AFF-RNN) that dynamically adjusts the forgetting factor in response to evolving error trends. The method was implemented in MATLAB and tested on simulated flight scenarios to assess its performance under different dynamic conditions. Results show that AFF-RNN achieved up to 25 % improvement in estimation accuracy compared to RNN-G. A stability analysis based on a normalized Lyapunov energy functional confirmed asymptotic convergence, ensuring reliability for safety-critical applications. AFF-RNN demonstrates consistent performance in aircraft parameter estimation with its flexibility, robustness, and theoretical assurances. It can also be used as a basis for future extensions to actual flight data in defense applications.
Bluff body flame holders are critical for liquid-fuelled ramjet combustors to ensure stable flame anchoring through the formation of a downstream recirculation zone. While Large Eddy Simulation (LES) provides high accuracy, its computational cost remains prohibitive for design iterations. This study evaluates the effectiveness of SST k-omega, RSM, and Scale-Adaptive Simulation (SAS) models in predicting unsteady flow features downstream of a Volvo triangular bluff body. The results demonstrate that while the SST k-omega model significantly over-predicts the recirculation zone length by nearly 15 %, the RSM model shows better agreement with experimental mean flow data. However, both traditional URANS models fail to resolve the high-frequency instabilities within the shear layer. In contrast, the SAS model successfully captures the dynamics of the flow, resolving the large-scale coherent structures and predicting a Strouhal number of 0.29, which aligns closely with experimental benchmarks. Furthermore, spectral analysis reveals that SAS captures the energy decay in the inertial subrange. The study concludes that SAS provides a superior balance of accuracy and computational efficiency, making it the most viable candidate for modelling unsteady wake dynamics in industrial combustor design.
Contained Detonation Chambers (CDCs) represent a significant advancement in the safe and environmentally responsible demilitarization of unserviceable munitions and unexploded ordnance. This quantitative review synthesizes current knowledge on CDC design, blast loading mechanics, material behaviour under extreme strain rates, and numerical modelling methodologies. The analysis shows that the structural response is primarily impulse-driven. Equivalent plastic strain and cumulative deformation are more reliable metrics for assessing reusability than peak stress or pressure. The use of normalized design parameters, such as the charge-to-volume ratio and thickness-to-radius ratio, enables scalable design and objective comparison across systems. The review demonstrates that dominant failure modes include incremental plastic deformation and low-cycle fatigue, which directly relate to reusability criteria. Equivalent plastic strain thresholds typically range from 0.1 % to 1.0 %. Best practices in numerical simulation require fully coupled fluid–structure interaction (FSI) and high-resolution mesh refinement. These are necessary to capture localized pressure amplification and momentum transfer. Material selection emphasizes high-strength low-alloy (HSLA) steels. The Johnson–Cook constitutive model effectively represents rate-dependent plasticity. This review also identifies research gaps in standardizing performance metrics, experimental validation, and integrated blast–thermal–off-gas modelling. These areas provide direction for future CDC development.
Automated ship detection and subsequent classification of maritime vessels operating within littoral zones are paramount for effectively safeguarding national and international maritime interests regarding ship traffic. Over recent years, different deep learning approaches have made notable progress in image analysis and classification. Still, it requires high processor power and memory space, which poses deployment challenges for such a system on satellites and edge devices. To address this, a lightweight deep learning network is designed to minimize computational parameters and complexity while retaining high recognition rate in ship classification task. The proposed method has incorporated the knowledge distillation technique, where a high-capacity teacher network guides the lightweight student model by transferring soft output distributions. Experimental results affirm that the proposed lightweight method has shown superior ship image classification performance compared to existing approaches, achieving the highest classification accuracy of 97.25% with least 2.35M parameters.
Autonomous traversability in unstructured roads demands automatic off-road track identification systems. This area has garnered significant research interest in recent times. The present study leverages the YOLOv8 transfer learning architecture and its variants for off-road track identification and segmentation to detect drivable tracks in off-road areas. The YOLOv8 model has been chosen for their advanced feature extraction capabilities. Currently, the off-road areas present the challenges of loosely organised class boundaries, non-uniform features, and a noisy environment. Thus, the present study provides an end-to-end strategy using an optimised YOLOv8l variant for terrain analysis and emergency decision-making, offering an alternative to conventional image processing and LiDAR-based approaches. The YOLOv8 model and its variants were evaluated and compared across both validation and inference datasets. Out of which the YOLOv8l model was tuned for selectable hyperparameters from the internal architecture to balance speed, computational efficiency, and detection precision across the large data set. The reliability of the optimised YOLOv8l model was evaluated on a real-time video feed and on unseen image scenes captured with a camera, and the model achieved an overall accuracy of more than 94 %. The model demonstrated its efficiency in snow regions for identifying unstructured tracks, showcasing its potential for autonomous military and emergency navigation operations in high-altitude conditions.
Aviation industry utilizes high end technology, critical materials, and components. Wide range of polymers are used in critical applications in aircraft. Helicopters utilise common as well as exotic polymers with cost always remaining secondary to safety and service life. Even though design, development and validation of the airframe, aero-engines and aggregates are done with stringent standards, failure do take place. The research work has evaluated attributability and root cause of the major failures due to polymeric components used in helicopters. Failures in helicopters are more acute than fixed wing aircraft due to their unique operational requirement in high altitude and extremely cold climatic conditions. Average service life of the helicopters and aeroengines are found to be intrinsic function of reliability of the polymeric materials. The research utilised case studies involving catastrophic failure due to various of rubber and PTFE. Fail safe design, black box development, utilization of newer materials like HNBR are potential ways to eliminate catastrophic failures.
Test Case Prioritization represents a fundamental strategy in software quality assurance that focuses on optimizing testing workflows through strategic sequencing of test execution. This review paper examines the evolving landscape of test case prioritization techniques, exploring methodologies and algorithms that have emerged to address the growing intricacies of modern software architectures. By synthesizing recent research findings, this paper identifies critical areas, emerging patterns and promising avenues for advancement in software testing domain.
This paper describes preparation and characterization of hydroxyl-terminated polyester (HTPE)-based energetic composites from high melting explosive (HMX) and polyester polyols by varying the amount of polymer binder content by cast cured technique. These composites possess several advantages, such as improved structural integrity, higher thermal stability, and a lower sensitivity when compared to TNT-based formulations. They are employed in various modern armaments, including bombs, missile warheads, and torpedoes.The HTPE prepolymers were cured with different curing agents to crosslink HTPE-based energetic composites. The effects of the different curatives on the chemical, thermal, mechanical and energetic properties were studied by employing different analytical instrumental techniques. The chemical stability assessed using a vacuum stability tester (VST) demonstrated significant high chemical stability. The hazard assessment in terms of the impact and friction sensitivity indicated that all energetic composites were less sensitive than those of pure HMX and slight varying in the sensitivity, depending on the curing agents. The thermal stability is either comparable or higher than that of the HMX. The compression strength of the HTPE-based composites was studied and found to be increasing with increasing the polymer content. The varying proportions of HMX and polyester were carried out to see the effect on the crosslinked elastomeric binders on the mechanical properties. The detonation velocity was obtained in the range of 7.0 km/s - 7.9 km/s with varying amount of the binder system from 10 % - 20 %.
This work presents the design, modeling, and simulation of an Electromechanical Actuator (EMA) for the Rustom-II UAV's Nose Landing Gear (NLG) steering. While Electromechanical Actuators have been widely studied for aerospace applications, a focused implementation for UAV NLG steering remains scarce in current literature, with most prior works emphasizing flight control or generic actuator design. To address this gap, we propose an integrated framework combining Brushless DC (BLDC) motor design, multi-stage planetary gearbox optimisation, and cascaded PID-based control loops. The BLDC motor was modelled for both interior and surface permanent magnet rotor-stator configurations, with performance evaluated using JMAG for electromagnetic and thermal analysis. Co-simulation using MSC Adams and MATLAB/Simulink enabled dynamic validation of the NLG steering system under operational loads. Results show that the 8 Pole-12 Slot IPM configuration delivers the required torque of 1.5-1.7 Nm at 670-1000 RPM with smooth speed -torque characteristics, while maintaining acceptable thermal limits. Compared to existing EMA designs, the proposed approach provides a compact, thermally robust solution tailored for UAV constraints such as weight, precision, and reliability. The findings demonstrate that the developed EMA can enhance manoeuvrability, reduce maintenance requirements, and improve dependability in Defence UAV applications.
Ablative nano composites were prepared by incorporating varying weight percentages of Carbon Black Nano particles (CBNP) into phenolic resin, then impregnated into rayon based carbon fabric material. CBNP was blended into the phenolic resin at loadings of 0 %, 2 %, 4 %, 6 %, and 8 % by weight using an electrical stirrer to ensure uniform dispersion of the nano particles. The laminates were fabricated using hand layup process and autoclave curing. Physical, mechanical, and ablative tests were conducted on the samples in accordance with current ASTM standards. The ablative tests were performed using an oxyacetylene test bed (OTB) in which the laminates were exposed to a flame of heat flux about 500 W/cm & sup2; for a duration of 60 sec. SEM results were analyzed on the burnt surfaces to understand the influence of CBNP nano fillers on ablation properties of all types of composites. It was observed that the composite specimen containing 4 % CBNP exhibited better ablation properties (reduced erosion rate, mass loss and back wall temperatures by 28 %, 9 % and 16 % respectively) when compared to control samples. This improvement can be attributed to the optimal interface formed between the fiber and resin at 4 wt % CBNP.
This paper introduces a novel pulse frequency modulation (PFM)-driven universal step-down converter tailored for high-efficiency, low-power IoT applications, with a strong emphasis on minimizing voltage ripple. The architecture incorporates a flexible Comparator in place of conventional static designs, thereby enhancing voltage gain, accelerating decision speed, and reducing dynamic power dissipation. A flexible OFF-control strategy is further employed to selectively deactivate auxiliary control circuits, significantly mitigating controller-induced losses. Fabricated in 0.32-gm CMOS technology, the converter achieves output ripple voltages consistently within 2.1-2.9 mV, while attaining a peak power efficiency of 95 %. These results underscore the converter's robustness and suitability for resource-constrained IoT environments. This innovative approach highlighting the potential of PFM-based control in addressing challenges associated with voltage regulation, power efficiency, and ripple reduction, making it an ideal solution for next-generation power management systems requiring compact, efficient, and reliable converters. This method demonstrates the effectiveness of Pulse Frequency Modulation (PFM)-based control in overcoming voltage regulation challenges, enhancing power efficiency, and minimizing output ripple. Its ability to deliver compact, reliable, and energy-efficient performance positions it as a highly suitable solution for next-generation power management systems in modern electronic defence applications.
Reverse Osmosis (RO) is a widely used technology in Defence operations to provide clean, potable water in remote and harsh environments. The water desalination systems available with the services experience operational discontinuities, treatment of a wide range of feedwater, which results in a significant enhancement of RO membrane fouling. Herein, the impact of fouling was studied using an in-house developed high-pressure desalination system. Experiments were conducted on two types of membrane modules, one with induced fouling using underground water (MM1) and a naturally fouled membrane module due to operational discontinuities over a period of 9 months (MM2). The performance of these modules was recorded before scaling, after scaling and, after scale removal in terms of per cent salt rejection, per cent recovery, product TDS, and product flux rate. It was observed that MMl fouled mainly due to inorganic scaling was easily regenerated using 2wt % citric acid cleaning solution, while MM2 fouled due to inorganic and biological agents was treated using tailored cleaning protocols, amongst which, sodium tripolyphosphate and sodium dodecylbenzene sulphonate mixture provided the best results. Overall, these findings underscore the necessity of proper mitigation strategies to combat fouling of RO membrane modules that are operated at a high pressure similar to commercial desalination systems.
Demand side management is a strategy to optimize and reduce electricity demand by urging users to alter their energy consumption patterns such that the load on the power grid is nominal. This work presents the development of a comprehensive smart energy management system for residential buildings, integrating sensors and controllers to provide full control over loads, along with a cloud-based setup that manages data logging and load schedule optimization. The proposed scheme optimizes loading schedules while reducing the Peak-to-Average Ratio and improving energy efficiency. Users are allowed to alter their optimized schedules to suit their needs, override automation when required, and view their energy consumption. This system also shows immense potential for cross-domain applications, such as in military field units. In off-grid military convoys, where diesel generators or portable batteries often provide electricity, effective energy management is crucial. The proposed scheme implements a novel objective function. It has achieved a 53 % reduction in the standard deviation, leading to a flatter load curve, ensuring better load distribution throughout the day and hence reducing pressure on the power grid. This work addresses various challenges and constraints in demand-side management, proposing an effective approach for optimized load scheduling that contributes to smarter and more efficient residential energy consumption.
In today's ever-changing IoT security environment, good Intrusion Detection Systems cannot be overemphasised. They are indispensable in shielding against all sorts of attacks and threats since no system is immune to threats. The four architectural layers of IoT are the application, processing, network, and perception layers. This research introduces an innovative deep learning model that encompasses previous solutions by classifying the attacks based on various layers of the IoT 4-layered Architecture. It is strategic because it makes organisations avoid being caught off guard by a breach by instead focusing on which areas can be targeted to guarantee strategic and preemptive safeguarding against emerging attacks. The proposed model integrates CNN, GRU, and ConvLSTM structures in a complex structure. The model undergoes rigorous evaluation on the NSL-KDD dataset, addressing binary and multiclass classification tasks. Experiment results are compared with five other hybrid models: CNN LSTM, LSTM GRU and finally, the Recurrent Neural Network with GRU. The experimental outcomes indicate that the suggested model outperformed the accustomed hybrid models applied for comparison concerning attack detection and classification in IoT networks. Collectively, the proposed integration improves the model in effectively capturing and analysing the intricate and latent characteristics of network traffic data, widely seen as one of the persistent challenges in IoT security.
This study explores the potential of recovering potable water from vehicle exhaust to serve as an emergency water source for Indian soldiers in desert regions. The research employs heat transfer applications, including heat exchangers and refrigeration systems, to collect condensate from exhaust emissions. To remove impurities from the collected water, it is first passed through a charcoal chamber. Further analysis focuses on evaluating the physical properties of water, such as pH, conductivity, turbidity, total dissolved solids, and hardness. Inductively coupled plasma optical emission spectrometry is utilized to measure elemental concentrations. Results indicate that with proper treatment, water recovered from vehicle exhaust can meet potable water standards. For instance, most of the test samples fell within WHO and BIS IS:10500 limits for parameters like pH, TDS (<500 mg/L), and metal contents. This innovative approach offers a viable solution for emergency water supply in arid environments, enhancing the efficiency and resilience of soldiers in challenging environmental conditions.
This paper presents the idea of isolation enhancement between radiating elements of a fifth-generation multiple antenna using a modified coupled line. Here, two radiating elements with dimensions of 19.4 mm x 25 mm are placed closely, with a minimal end-to-end spacing of 0.072 X. Mutual coupling reduction is done by placing a coupled line modified with meandering along with triangular grooves. Antenna resonates at 3.69 GHz, suitable for n77 5G NR band with an isolation enhancement of 33.06 dB. Isolation is verified by calculating the voltage between the radiating and isolating elements. MIMO performance parameters like envelope correlation coefficient (ECC) value of 0.001 and diversity gain (DG) value of 9.97 dB are achieved for the proposed system. The proposed system operates in the Sub 6 GHz 5G band and is hence suitable for 5G applications. Experimental results match well with simulation results.
A novel slot antenna loaded with strips suitable for multiband operations, is demonstrated in this article. A basic slot antenna is designed which is loaded with plus and cross shaped strips for resonating at multiple bands. The antenna resonates at 2.4 GHz, 6.5 GHz, and 9 GHz with a gain of 3.51 dBi, 3.56 dBi and 7.7 dBi. The antenna is fabricated on Rogers RT/Duroid 5880 material of size 50 & times;50 & times;1.6 mm3. The accuracy between the theoretical study and simulated outcome is then confirmed by manufacturing and testing the proposed double layered slot antenna. The measured outcomes, which largely match the simulated outcomes, demonstrate that the designed antenna can function in three bands of 680 MHz centered at 2.4 GHz, 250 MHz centered at 6.5 GHz and 540 MHz centered at 9 GHz. The suggested design is compatible with Radar systems and Wi-Fi wireless connection, and it provides reliable omnidirectional and bidirectional patterns.
A Substrate Integrated Waveguide (SIW) antenna array is presented in this article for use in active Phased Array antennas (PAA) with Tile based architectures. A modified E-shaped patch antenna element with calibration lines embedded on the same layer is designed and realized. Rogers RT Duroid 5880LZ Substrate with thickness 0.036 a.degrees is used for realizing the array. The calibration lines are printed parallel to the E-plane of the array in between the SIWs and can be used for fast online calibration. The 64-element antenna array has been designed to operate in X-Band and has been fabricated and tested by exciting the center element. The measured result confirms an impedance bandwidth of 7.95 %, cross-polarization better than 20 dB for the center element of the array. The Active Element Pattern (AEP) confirms the peak gain of 5.24 dB over the operation band. The array has a wide scan capability of +/- 50 degrees in E-and H-planes respectively with a broadside gain of 23 dBi. The low-profile of the array with inbuilt calibration lines enables its application in radars with tile-based architecture.
To overcome the limitations of conventional radar target recognition methods, this paper proposes a novel approach that integrates cyclic spectrum slicing with deep learning techniques. First, the cyclic spectrum of the radar signal is computed using the time smoothing method, enabling a theoretical analysis of the distinguishing features present in the target’s cyclic spectrum slice. The resulting slice spectrum is then fed into an improved Deep Convolutional Generative Adversarial Network (DCGAN) model, which employs a Wasserstein GAN with Gradient Penalty (WGAN-GP) loss function for data augmentation. Subsequently, an enhanced VGGNet architecture—incorporating Global Average Pooling and Parametric Rectified Linear Unit (PReLU) activation functions—is utilized to automatically extract features that capture the target’s cyclic stability. Experimental results on aircraft target recognition demonstrate that the proposed method achieves outstanding average accuracies of 98.46 % and 98.40 % for different cyclic spectrum slices, significantly outperforming traditional methods by more than 3 %. The findings highlight the exceptional capability of the cyclic spectrum in revealing the intrinsic properties of target signals, along with its strong suppression of noise and clutter. Moreover, the integration of deep learning techniques has substantially improved target recognition accuracy, offering an effective solution for automatic target recognition under conditions of limited samples and low signal-to-noise ratios.
The information security, which has the utmost importance in present scenario, essentially protects system resources to guarantee CIA (Confidentiality, Integrity and Availability). And the systems meant to provide security are commonly known as Encryption Devices. The encryption devices basically count on cryptographic algorithms which turns the plain text into cipher text. This paper discusses an automation technique for the verification of cryptographic algorithm applications implemented in HDL (Hardware Description Language) for FPGA based encryption systems. Verification ensures that cryptographic algorithm is faithfully implemented in the targeted FPGA as per the intended design. However, the verification of entire FPGA application design has become the challenge due to the increased complexity of modern encryption systems. The developed technique overcomes the challenges in the verification process of cryptographic algorithm HDL application. The automated verification approach overrules the demand of manual efforts, enhances the depth of verification by ensuring exhaustive coverage, ensures high level of confidence and most importantly fast convergence of verification signoff. The proposed technique has auto scalable and highly configurable architecture. It is evolved and tested for a proprietary stream cipher, it is undoubtedly applicable for any stream ciphers in general and can be adopted for block ciphers too with minimal efforts.
