Rajiv Gandhi University of Knowledge Technologies

Abstract The adoption of energy-efficient and intelligent control systems is critical for modern LED-based street lighting solutions. Traditional control methods, while effective, often struggle to adapt to non-linearities and dynamic changes in system parameters. This paper presents the design and implementation of a novel sensorless controller for boost converter driven DC LED Street lights, leveraging deep learning to achieve accurate and robust voltage regulation without a physical voltage sensor. A Deep Neural Network (DNN) is trained to estimate the output voltage of the system using measurable parameters such as input voltage, load current, and duty cycle. This estimated voltage is then utilized in a control loop to maintain stable operation under varying load and environmental conditions. The DNN model is trained on simulated and experimental data representing a wide range of operating conditions, including load variations. The controller is implemented on DSP controller platform and validated on a laboratory prototype of 500W DC boost converter-based LED driver. Performance evaluation is demonstrated using simulation and experimental results. This work provides a scalable and cost-effective solution for modern street lighting systems, paving the way for smarter and more sustainable urban lighting infrastructure.

Accurate segmentation of water bodies from satellite imagery is essential for sustainable water resource management, flood monitoring, and hydrological modeling. In recent years, many deep convolutional neural network (DCNN) based techniques have been developed for the purpose of water bodies segmentation. These models often suffer from issues such as class imbalance, poor land-water contrast in urban environments, and loss of fine boundary details, although they have significantly improved segmentation accuracy over traditional methods. In order to improve spatial feature refinement, this study proposes a modified U-Net architecture that incorporates residual connections into the skip connections, followed by a Dual Attention Block (DAB). Furthermore, by combining features from several receptive fields, a Multi-Scale Fusion Block (MSFB) is incorporated at the bottleneck to improve contextual understanding. The model was trained on Sentinel-2 satellite images and their corresponding segmentation masks obtained from the public dataset. Compared to traditional DCNN architectures, the proposed U-Net based model provides better qualitative and quantitative results and captures small and fragmented water features with greater robustness to spatial complexity and environmental variability.

Polycystic Ovary Syndrome (PCOS) is a complex endocrine disorder affecting women of reproductive age, and early detection is critical for effective management. In this study, we performed data preprocessing, feature extraction, and used clustering algorithms to improve PCOS detection. The dataset underwent resizing, gray scale conversion and noise removal during preprocessing to enhance data quality. Features were extracted using a pre-trained ResNet50 convolutional neural network, leveraging its deep hierarchical representations for robust feature learning. These extracted features were then utilized by clustering algorithms, including DBSCAN, OPTICS, K-means and BIRCH, to group patient samples and identify underlying patterns without explicit labels. Experimental results demonstrated that the combination of ResNet50-based feature extraction and clustering achieved better separability between PCOS and non-PCOS groups compared to raw data clustering. The findings highlight the potential of coupling deep feature extraction with unsupervised learning for PCOS detection and suggest a promising direction for developing scalable, automated, and label-independent diagnostic support systems.

A new library of α-aminophosphonate derivatives (9a–j) was developed utilizing a Kabachnik–Fields multicomponent method with microwave assistance. Prior to synthesis, their pharmacokinetic characteristics and inhibitory interactions with the enzymes α-amylase and α-glucosidase were predicted using molecular docking simulations and in silico ADMET evaluation. Spectroscopy was used to check the structural integrity. The synthesized compounds were examined in vitro for inhibitory effects against two enzymes associated with type 2 diabetes. Compound 9e inhibited α-glucosidase the most (IC50 = 88.89 μg/mL), while compound 9b inhibited both enzymes (IC50 = 2.6 μg/mL for α-glucosidase and 102.36 μg/mL for α-amylase), similar to acarbose. SAR analysis found that electron-donating and heterocyclic groups increased activity, whereas bulky or electron-withdrawing substituents reduced efficacy. Statistical validation with ANOVA and Tukey’s HSD test revealed significant variations in enzyme inhibition across the compound set. These findings suggest that the synthesized compounds serve as promising candidates for the development of dual-action anti-diabetic agents.

Background: Congenital heart anomalies (CHA) remain one of the leading causes of childhood morbidity and mortality worldwide, with significant implications for long-term health outcomes and healthcare systems. Method: Utilizing global burden of disease 2021 study standardized methodology we estimated CHA burden for all 27 European Union countries from1990-2021. Age-standardized rates of deaths, prevalence, and years lived with disability (YLDs) were analyzed, and annualized percentage changes (APC) were calculated for multiple time intervals to assess temporal and sex-specific trends. Result: From 1990 to 2000, the APC in deaths due to CHA declined significantly by -5.72%, followed by a continued but less steep decline of -3.90% from 2000 to 2021. In contrast, the prevalence showed a slight annual decrease of -0.12% during 1990–2000, which reversed to a modest annual increase of 0.19% in the period 2000–2021. YLDs demonstrated a similar trend, with an annual decline of -0.58% from 1990 to 2000 and a slight increase of 0.12% annually from 2000 to 2021. The age-standardized death rate (ASMR) declined annually by -5.10% from 1990 to 2000 and -4.00% from 2000 to 2021, while the age-standardized prevalence rate (ASPR) remained largely stable during 1990–2000 (0.006% per year) and rose slightly thereafter (0.11% per year); similarly, YLDs (ASYLDsR) showed a mild decline in the first decade (-0.17%) followed by a gradual annual increase of 0.19% from 2000 to 2021. Nationally, the highest APC in ASPR observed in Austria by 0.66%, followed by Ireland 0.23%, Portugal 0.22%, Croatia 0.19%, Malta 0.16%. From 2000 to 2010, the ASMR declined annually by -4.30% in males and -4.58% in females, with a slower decline observed from 2010 to 2021 (-3.48% in males, -3.80% in females). The prevalence rate showed a modest annual increase in males (0.28%) and a slight decline in females (-0.05%) during 2000–2010, followed by a small rise in both sexes from 2010 to 2021 (0.02% in males, 0.18% in females). YLDs remained stable in females (-0.007%) and increased marginally in males (0.23%) from 2010 to 2021. Conclusion: While the mortality due to CHA has steadily declined across 3 decades and sexes, both prevalence and disability burden have shown a gradual upward trend since 2000, particularly among males. These findings highlight a shifting burden from fatal to non-fatal outcomes, underscoring the need for long-term care strategies and quality-of-life interventions.

Background: Chronic kidney disease(CKD) is a growing global health challenge, with hypertension being a leading cause of renal dysfunction across populations. A major yet underrecognized complication of hypertensive CKD (HTN-CKD) is anemia, which not only worsens renal outcomes but also amplifies cardiovascular risk through mechanisms like left ventricular hypertrophy and myocardial strain. Understanding its burden is critical, as anemia in this setting accelerates disease progression and contributes significantly to cardiovascular morbidity and mortality. Method: We estimated the global and regional burden of anemia attributable to HTN-CKD in 204 countries and its terriotories from 1990-2021 using standardized Global Burden of Disease Study 2021. Estimates of prevalence and years lived with disability(YLDs) were generated using DisMod-MR 2.1, a Bayesian meta-regression tool that integrates data from population-based surveys, cohort studies, hospital records, and vital registration systems. Age-standardized rates were calculated to facilitate comparisons across time and regions, and annualized percentage changes(APCs) were computed using log-linear regression. Results: Between 1990 and 2021, the global prevalence of anemia attributable to HTN-CKD more than doubled, rising from 2.6 million(95% UI: 2.4–2.9 million) to 5.6 million(5.2–6.1 million), with an APC of +2.46%. Correspondingly, YLDs increased from 74,819 (49,512–106,554) to 138,672 (92,028–198,889), reflecting an APC of +2.01%. The highest regional increase in prevalence was observed in South Asia (APC: +3.16%), followed by Latin America and the Caribbean (+2.94%), Sub-Saharan Africa (+2.80%), and North Africa and the Middle East (+2.71%). The greatest rise in age-standardized prevalence rate was seen in Fiji (APC: +0.80%), followed by the United States (APC: +0.74%), signaling a growing concern even in high-resource settings like the USA. Among adults aged 70 years and above in middle and low-middle SDI regions, both prevalence and YLDs showed annualized increases exceeding 2%, whereas younger age groups in high SDI regions exhibited stable or declining trends. Conclusion: The global burden of anemia due to HTN-CKD has escalated, with the sharpest increases concentrated in low- and middle-income regions where hypertension detection and management remain suboptimal. The burden is especially pronounced among older adults and men, accentuating widening disparities by age, sex, and socioeconomic status.

Heart failure (HF) is a growing global health concern, and cirrhosis is an increasingly recognized non-cardiac contributor through mechanisms like volume overload, systemic inflammation, and cirrhotic cardiomyopathy. As the global prevalence of liver disease rises, especially from nonalcoholic fatty liver disease (NAFLD), alcohol use, and viral hepatitis, understanding the burden of HF due to cirrhosis is essential. We estimated the non-fatal burden of HF attributable to cirrhosis across five etiologies—NAFLD, alcohol use, hepatitis C, hepatitis B, and other causes—in 204 countries and territories from 1990 to 2021 using data from the Global Burden of Disease Study 2021. Non-fatal burden was estimated by applying standardized severity distributions and disability weights to the prevalence of each cause to quantify impairment. Age-standardized prevalence rates (ASPRs) and annualized percentage changes (APCs) were calculated by SDI, age, and sex. Future prevalence to 2050 was projected using a log-linear regression model. Between 1990 and 2021, total prevalence counts rose substantially across all cirrhosis subtypes. NAFLD-related HF increased from 9,202 (95% UI: 6,417–12,641) to 30,092 (20,452–42,009); alcohol-related from 35,701 (27,666–45,427) to 82,258 (61,635–108,803); hepatitis C from 47,484 (37,044–60,873) to 92,079 (70,272–120,432); hepatitis B from 59,974 (47,824–74,529) to 112,111 (87,717–142,335); and other causes from 99,977 (77,246–129,309) to 133,311 (102,461–169,414). The highest ASPR increase was seen in low SDI regions (APC: 2.49%), followed by high SDI (2.24%), middle (1.73%), high-middle (1.67%), and low-middle SDI (1.60%) regions. In 2021, children aged 5–9 years exhibited the highest burden within the under-20 group (26,749 cases), while adults aged 65–69 years had the peak burden among older individuals (51,379 cases). Females showed a marginally higher annual growth in total prevalence (APC: 0.79%) and ASPR (0.04%) compared to males (0.78% and 0.01%, respectively). Projections estimate the prevalence of heart failure due to cirrhosis will increase to 788,332 cases globally by 2050. HF attributable to cirrhosis is a growing global health concern with rising prevalence across all etiologies, disproportionately impacting low-SDI regions and older adults. The projected surge by 2050 underscores an urgent need for integrated liver-cardiac care, early detection, and policy interventions targeting modifiable liver disease risk factors.

This paper introduces a new approach to improving communication for deaf and hard-of-hearing individuals by developing a multilingual voice and text–to–sign language translator. Unlike existing tools that often lack support for multiple languages, struggle with real-time accuracy, or only handle text or voice separately, our solution combines both voice and text inputs in one system. The framework uses advanced AI technologies, including speech recognition (ASR), natural language processing (NLP) for context-aware understanding, and deep learning to generate realistic sign language. It is designed to support different sign languages (such as ASL and ISL) through specialized datasets, while addressing common challenges like limited vocabularies, robotic or unnatural avatars, and weak handling of continuous signing. The results show that this integrated AI model can create a more natural and inclusive communication experience, with the potential to promote social inclusion, expand educational opportunities, and bring economic benefits.