Vignan's Institute of Information Technology

The security and stability of modern power systems rely heavily on timely and accurate detection of anomalies, including cyberattacks. Traditional anomaly detection approaches often struggle to capture complex interdependencies among system components, resulting in decreased detection performance. In this study, we propose a novel deep learning-based method for detecting anomalous events in power networks. The proposed model leverages historical and real-time measurement data to identify deviations that indicate potential false data injection attacks. Extensive experiments were conducted on standard IEEE test systems, incorporating adversarially injected cyberattack signals to evaluate the robustness of the approach. The results demonstrate superior performance in accuracy, precision, and recall compared with conventional autoencoder-based models. This method provides a reliable tool for enhancing the cybersecurity and operational reliability of power systems.

The integration of sustainable waste-derived reinforcements into polymer composites is an effective strategy for improving performance while reducing environmental impact. In this study, rice husk biochar was investigated as a particulate reinforcement for epoxy composites, and its influence on mechanical, tribological, and thermal behaviour was systematically evaluated. Epoxy composites containing different weight fractions of rice husk biochar were fabricated and characterised. Tensile strength increased from 22.7 MPa for neat epoxy to 29.2 MPa at 9 wt.% biochar, accompanied by an increase in elongation at break from 0.8% to 1.31%, indicating improved stress transfer and reduced brittleness. Flexural strength similarly improved from 58 MPa to 70.1 MPa, confirming enhanced resistance to bending-induced failure. Fracture-surface analysis revealed suppression of cleavage-dominated river patterns and increased crack deflection in biochar-reinforced composites. Dry sliding wear analysis showed that intermediate biochar content increased wear mass loss (0.76 mg at 6 wt.%) due to particle pull-out and third-body abrasion, whereas higher filler loading promoted more stable surface interaction. Thermogravimetric analysis demonstrated improved thermal stability of biochar-reinforced epoxy composites at elevated temperatures, particularly in the 400 degrees C-550 degrees C range, attributed to char-mediated thermal shielding. Overall, the results demonstrate that rice husk biochar provides multifunctional enhancement of epoxy composites by improving mechanical performance and high-temperature thermal resistance, while introducing content-dependent tribological effects, highlighting its potential as a sustainable reinforcement for epoxy-based non-load bearing structural applications.

Polymer-based composites have gained prominence in tribological applications due to their lightweight nature, tunable properties, and multifunctional potential. However, existing reviews largely report performance improvements without systematically addressing contradictory trends, testing variability, and emerging manufacturing routes. This review analyses friction and wear mechanisms in fibre-reinforced and particle-reinforced polymers, surface coatings, and additively manufactured polymer composites. Key mechanisms, including load transfer, transfer film formation, thermal dissipation, and interfacial effects, are critically synthesised across thermoset and thermoplastic systems. Representative performance trends are discussed to highlight the influence of reinforcement type, processing route, and operating conditions, along with limitations in current tribological testing practices and the need for standardisation. By integrating mechanistic understanding with comparative performance and future research priorities, this review provides guidance for the design and evaluation of polymer composites in automotive, aerospace, marine, and biomedical tribological applications.

The current paper proposes a support system to increasing the stability of autonomous vehicles by using the combination of reinforcement learning (RL) and bio-inspired control strategy. In particular, we suggest to utilize genetic algorithms to learn optimal control to RL so that the system is able to learn in changing road conditions and disturbances. The genetic algorithm aids in the search of the solution space which seeks the best performing policy to enhance the vehicle stability control. We use the CARLA simulator as an instrument to evaluate and verify the developed approach in a range of driving situations, in sharp turns and slippery roads. The stability aspect of our approach is much better as the response times and change of control was smoother than in the traditional methods. The findings suggest that RL coupled with bio-inspired optimization procedures provides a resilient adaptive solution to obtaining stability improvement in autonomous vehicles, and thus it is suitable to real life situations in autonomous driving systems.

Future 6G wireless networks require communication paradigms that go beyond traditional bit-based transmission toward intelligent, task-oriented information exchange. This paper proposes a Digital Twin–aided semantic communication framework that integrates knowledge-driven semantic inference, end-to-end deep learning–based physical-layer optimization, and cognitive multi-objective optimization. Unlike fully data-driven semantic systems, the proposed semantic transmission is formulated as a statistical inference problem, where task-relevant information is modeled probabilistically and optimized under channel uncertainty. An end-to-end learnable physical layer minimizes reconstruction distortion while maximizing semantic efficiency. Simulation results demonstrate that the proposed model reduces mean squared error from 0.185 to 0.132 at 0 dB SNR and 0.072 to 0.038 at 10 dB SNR compared to conventional systems. Semantic accuracy reaches 74.5% at low SNR, 98.6% at high SNR, with a semantic efficiency of 0.87. Furthermore, Digital Twin–assisted cognitive optimization reduces average latency from 42.6 ms to 27.8 ms, energy consumption from 18.9 mJ to 11.3 mJ, and improves adaptability to 0.86. The proposed framework provides an interpretable, scalable, and AI-native foundation for next-generation 6G communication systems.