Volume 14 | Issue 5 |

 Abstract

Smart grids are modern electrical networks that integrate digital technologies, smart meters, and communication systems to efficiently monitor and manage electricity generation, transmission, and distribution. However, increasing reliance on digital infrastructure makes smart grids vulnerable to cyber threats such as False Data Injection (FDI) attacks and energy theft, which manipulate meter readings or sensor data leading to incorrect billing, operational instability, and potential power system failures. This paper proposes a Convolutional Long Short-Term Memory (ConvLSTM) based optimized model for detecting anomalies in smart grid systems. The model combines the strengths of Convolutional Neural Networks (CNN) for extracting spatial features and Long Short-Term Memory (LSTM) networks for learning temporal patterns from time-series smart meter data. The dataset employed consists of smart grid energy consumption data containing both normal and anomalous patterns. Data preprocessing techniques including handling missing values, outlier reduction, normalization using standard scaling, and class balancing via SMOTE are applied to improve model performance. The ConvLSTM model is trained and evaluated using standard performance metrics. Experimental results demonstrate that the proposed model achieves an accuracy of 78.92%, demonstrating reliable anomaly detection and improved identification of abnormal smart grid behaviors to help utility providers enhance grid security.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Anomaly Detection, Smart Grid, ConvLSTM, Deep Learning, False Data Injection, Energy Theft, Time-Series Analysis, Cyber Security.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Global supply chains are increasingly exposed to disruptions caused by geopolitical tensions, pandemics, cyber threats, climate events, and regulatory uncertainty. While traditional supply chain strategies have emphasized cost efficiency and lean operations, recent crises have highlighted the need for resilience--the capacity to anticipate, absorb, respond to, and recover from disruptions. At the same time, Artificial Intelligence (AI) has enhanced forecasting, visibility, and real-time decision support through machine learning, predictive analytics, and digital twins. However, the impact of AI depends not only on technical capability but also on how supply chain partners respond strategically to AI-generated information. This paper develops a conceptual framework that integrates AI, John Nash's game theory, and supply chain resilience. Suppliers and manufacturers are modeled as rational actors choosing between cooperation (information sharing and joint risk mitigation) and non-cooperation (independent decision-making). The framework proposes that AI reduces information asymmetry, improves transparency, and reshapes payoff structures, making cooperative equilibria more attractive and stable. These equilibria enhance preparedness, response speed, recovery effectiveness, and organizational learning. The study contributes by linking digital intelligence, strategic behavior, and resilience in a unified theoretical model. It also offers practical guidance for managers seeking to use AI as a strategic coordination mechanism rather than merely an operational tool. The framework provides a foundation for future empirical, simulation-based, and analytical research on resilient and intelligent supply networks.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Artificial Intelligence; Game Theory; Supply Chain Resilience; Disruption Risk; Strategic Decision-Making; Information Asymmetry; Nash Equilibrium; Collaborative Supply Chains

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

This research focuses on the development of an advanced system for accident detection and automatic ambulance alert to enhance emergency response efficiency. With the increasing frequency of road accidents, rapid detection and immediate response are critical for minimizing casualties. The proposed system utilizes state-of-the-art sensors and image processing techniques to accurately identify and analyze accident scenarios in real-time. Key components of the system include onboard sensors such as accelerometers and gyroscopes . These sensors continuously monitor vehicle dynamics and surroundings, enabling the system to detect sudden changes indicative of an accident. Once an accident is confirmed, the system automatically triggers an alert to emergency services, providing crucial information about the incident's location and severity. Furthermore, the system incorporates geolocation technology to precisely pinpoint the accident site, facilitating quicker ambulance dispatch. The automated ambulance alert aims to reduce response times, ensuring timely medical assistance for accident victims. This innovative approach not only streamlines emergency services but also contributes to the overall improvement of road safety. The study demonstrates the feasibility and effectiveness of the proposed system through simulations and real-world testing, showcasing its potential to significantly impact emergency response outcomes in accident scenarios.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Accident Detection and Automatic Ambulance Alert in the project is described in the diagram below. Here is a brief description of each of them and how they all interact with one another

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

This work presents a system for grouping and analysing music using machine learning and user-focused design. The system converts audio signals into numerical features such as MFCC and STFT, then groups similar tracks using clustering methods. It combines multiple clustering approaches to improve grouping stability and accuracy. The design also focuses on user needs through iterative development and feedback. Results show better grouping quality and improved usability. The system supports applications like recommendation systems and audio analysis.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Keywords : Music Clustering, Feature Extraction (MFCC, STFT), Hierarchical Classification, Consensus Clustering, User-Centered Design (UCD), Multi-View Learning, Music Recommendation Systems.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

This paper explores the profound thematic and structural parallels between Masuji Ono in An Artist of the Floating World and Stevens in The Remains of the Day. Both protagonists function as high-level stewards of crumbling ideologies--specifically the imperialist fervour of Sh?wa-era Japan and the rigid class hierarchy of the British landed gentry. By examining their roles as unreliable narrators and "victims" of their own professional devotion, this study argues that both men are casualties of a "falling structural world." In this context, the moral frameworks they served have not only collapsed physically and politically but have been retroactively condemned by the very societies they aimed to improve. Their tragedy lies in the irreconcilable gap between their subjective sense of "dignity" and the objective historical reality of their complicity.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Kazuo Ishiguro, Masuji Ono, Stevens, unreliable narration, moral ambiguity, historical complicity, social isolation, guilt, denial, emotional suppression

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

With the rapid advancement of cloud computing and the widespread adoption of containerized applications, managing high-concurrency workloads efficiently has become a significant challenge in modern distributed systems. Traditional auto-scaling techniques, particularly reactive approaches such as threshold-based scaling, often fail to respond promptly to sudden and unpredictable workload variations, resulting in resource over-provisioning or under-provisioning and degraded system performance. To address these limitations, this paper proposes a trend detection-based auto-scaling method that integrates Long Short-Term Memory (LSTM)-based workload prediction with a trend detection mechanism. A cooldown strategy is also implemented to prevent frequent and unnecessary scaling operations, ensuring system stability. The proposed approach effectively combines prediction, trend awareness, and controlled execution to optimize resource utilization while maintaining performance under dynamic workload conditions in containerized environments.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Auto-scaling, LSTM, Trend Detection, Kubernetes, Containers, High-Concurrency, Cloud Computing, Workload Prediction, Pod Allocation.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

In today's rapidly advancing technological world, efficient battery management is essential for applications such as electric vehicles, renewable energy systems, and portable electronics. A Battery Management System (BMS) ensures safe and reliable operation by monitoring key parameters like voltage, current, and temperature; however, existing systems often lack features such as user-defined charging limits, real-time State of Charge (SoC) and State of Health (SoH) monitoring, and effective fault protection. This paper presents the design and implementation of an improved BMS that incorporates a smart charging limiter, real-time monitoring, and MOSFET-based protection mechanisms. The system utilizes an analog front-end and a microcontroller to continuously monitor battery parameters and automatically protect against over-voltage, under-voltage, over-current, and overheating conditions. Additionally, the system provides monitoring of key parameters such as voltage, current, temperature, SoC, and SoH through a mobile application using the Blynk platform. The proposed system enhances battery safety, improves performance, and extends battery lifespan, making it suitable for modern energy storage applications.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

1. Battery Management System (BMS) 2. Lithium-Ion Battery 3. State of Charge (SoC) 4. State of Health (SoH) 5. Real-Time Monitoring 6. Smart Charging Limiter 7. Battery Safety 8. BQ76920 9. ESP32 Microcontroller 10. MOSFET Protection 11. CAN Communication 12. Blynk Platform 13. IoT-Based Monitoring 14. Cell Balancing 15. Protection Circuitry

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

The global digital economy is undergoing a structural transformation in which consumers are progressively transitioning from ownership rights to perpetual licensees paying recurring rents for digital goods and services. This paper explores the "Techno-Feudalism" framework advanced by Yanis Varoufakis, contrasting it against the "Platform Capitalism" counter-narrative, to examine how digital platforms in India extract cloud rent from consumers and businesses through algorithmically governed ecosystems. Employing a mixed-methods design combining secondary data analysis of global subscription economy trends with an exploratory pilot survey of urban Indian digital consumers (n=100), the study investigates the economic drivers compelling the transition from ownership to subscription models, the theoretical validity of digital rentier frameworks, and the psychological impact of subscription proliferation. Key findings reveal that 86% of the primarily urban, Gen-Z respondents prefer perpetual ownership over subscription access, yet 86% currently pay for at least one subscription service. This reflects a structural preference-reality dissonance driven by ecosystem lock-in, choice architecture, and the systematic elimination of buy-once alternatives. Platform entrapment is widespread, with 86% of respondents feeling trapped within digital ecosystems, and the average perceived ownership score standing at only 2.43 out of 5. The paper situates these findings within the theoretical debates between Technofeudalism and late-stage capitalism, while examining India's stalled Digital Competition Bill 2024 and institutional capacity constraints at the Competition Commission of India as of early 2026. The research concludes that without robust ex-ante legislative intervention and institutional capacity building, the Indian digital economy risks consolidating into a rentier system structurally constraining consumer economic agency.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Techno-feudalism, platform capitalism, subscription economy, digital ownership, cloud capital, consumer behaviour, Digital Competition Bill India, digital fatigue.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Anomaly detection in network traffic plays a critical role in ensuring network security, as unusual patterns may indicate cyberattacks, intrusions, or system failures. With the rapid growth of modern networks, the volume and complexity of traffic data have increased significantly, making traditional rule-based and statistical techniques less effective in identifying abnormal behaviour. This paper explores the use of machine learning techniques for efficient and accurate anomaly detection in network traffic. The proposed approach focuses on analysing network traffic characteristics and leveraging data-driven models to automatically learn normal and abnormal patterns. Machine learning algorithms including Isolation Forest, Naive Bayes, XGBoost, LightGBM, and Support Vector Machine (SVM) are employed to process large-scale datasets, enabling the detection of subtle and complex anomalies that are difficult to identify using conventional methods. The system incorporates data preprocessing using the KDDCup99 dataset, feature extraction via Principal Component Analysis (PCA), and class balancing through the Synthetic Minority Over-sampling Technique (SMOTE). The effectiveness of the proposed method is evaluated using standard performance metrics such as accuracy, precision, recall, and F1-score. The results demonstrate that machine learning-based approaches provide a scalable and reliable solution for real-time anomaly detection in modern network environments.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Anomaly Detection, Network Traffic, Machine Learning, KDDCup99, PCA, SMOTE, XGBoost, LightGBM, Isolation Forest, Intrusion Detection.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

In today's digital era, the secure storage and transmission of sensitive documents is a significant challenge. Manual record-keeping and traditional password-based systems are prone to threats such as unauthorized access, theft, and data manipulation. To resolve these concerns, this research proposes a secure Digital Locker System that enables users to store, upload, encrypt, and manage personal documents using AES-256 encryption with Fernet implementation and OTP-based authentication through Twilio. The system is developed using the Flask framework and provides a user-friendly interface resembling a physical locker. The Digital Locker ensures confidentiality, integrity, and high security of sensitive documents such as ID proofs, certificates, medical reports, and legal papers. This project demonstrates a scalable, secure, and user-centric digital document vault for individuals and organizations.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Digital Locker, Encryption, AES-256, OTP Verification, Flask, Secure Storage, Authentication

  License

Creative Commons Attribution 4.0 and The Open Definition