Bikash Chandra Sahana

Malaria is caused by Plasmodium parasites transmitted through the bites of infected female Anopheles mosquitoes, and remains a significant global health issue. Accurate and timely diagnosis of malaria is essential for effective treatment, yet conventional microscopic examination remains prone to human error and delay. The proposed study introduces an innovative multi-model ensemble approach that integrates the pre-trained backbone models MobileNetV2 and EfficientNetB0 networks to enhance malaria parasite detection performance. The preprocessing steps were applied to the images for outlier detection and removal before being fed into the proposed model to achieve an impressive accuracy of 99%. The proposed ensemble model outperformed various pre-trained Deep Learning (DL) models and state-of-the-art techniques. The model was evaluated using thin-blood smear images from the publicly available NIH Malaria dataset with 27,558 images, demonstrating its potential for application in clinical settings and contributing to more efficient and accurate malaria parasite detection using blood cell smear images. Further, the proposed model was deployed in a Python-based web application to improve user interaction.

The prediction of the duration of the occurrence of low-visibility events in a calendar day is a difficult process because of the complex and chaotic mechanisms of the onset and dissipation of the low-visibility events. However, it is most useful for the operation of airport services (scheduling of aircraft, optimal operations of the airports) and the planning of any activities (travel, tourism, agriculture, etc.). This research tries to build the best dynamic weighted ensemble of the best combination of base machine learning( ML) models (Light Gradient Boosting Machine (Light GBM), Random Forest (RF), and Support Vector Regression (SVR)) to accurately nearcast the duration of low-visibility events( fog (surface visbility<1000 m) and dense fog (surface visbility<200 m)) for a calendar day based on the initial conditions of 1500 UTC (Universal Time Co-Ordinate). Conditions such as surface meteorological parameters (air temperature, dew point temperature, relative humidity, wind (every 3 hours), rainfall (daily), and sunshine (daily)) and upper air meteorological parameters (wind, temperature, and relative humidity of 1000, 925, and 850 hPa (every 3 hours)) were taken into account to find the best set of explanatory factors for the accurate nearcasting of the duration of the low visibility events. The Pearson correlation coefficient and Spearman's rank correlation coefficient were used to choose the final set of model explanatory variables. The datasets were thoroughly examined using supervised ML algorithms at the various stages of training, testing, modelling, and cross-validation. All the best combination models' accuracy was evaluated and compared using performance measures, namely MAE (mean absolute error), RMSE (root mean square error), and R2( R squared error). Based on the coefficient of determination( R2), it can be observed that the suggested dynamic weighted ensemble model exhibits the best level of prediction accuracy, specifically 0.89 and 0.88 for the duration of fog and dense fog for a given lead time of a day. This surpasses the accuracy of LightGBM (0.79), RF (0.78), and SVR (0.76) for the prediction of the duration of fog. Therefore, this study highlights the potential of machine learning in facilitating the advancement of automation in airport scheduling and optimising the operations of airports, specifically in the fog-prone Indo-Gangetic Plains(IGP).

In beyond 5G (B5G) wireless communication, the non-orthogonal multiple access (NOMA) technique and reflecting reconfigurable intelligent surface (RIS) have been projected as two promising key technologies for vehicle-to-anything (V2X) communication. This paper considers a RIS-aided V2V network model employing power-domain (PD) NOMA. The scenario focuses on a deep-fade dense urban environment, where no line-of-sight (NLoS) path is available for direct communication between the vehicles. Due to the mobile nature of the vehicles, imperfect channel state information (CSI) has been considered in this work. Two communicating RIS, one located near the transmitting vehicle and another near the receiver vehicle, have been used in the system model. To assess the performance of the proposed model, the closed-form expressions of the outage probability, ergodic capacity and energy efficiency have been derived. To obtain further insights, the number of RIS elements is varied to obtain the impact of spatial diversity on each parameter. The obtained simulation results are further compared with the traditional Orthogonal Multiple Access (OMA) and pre-existing literature work for benchmark comparison. The result suggests that RIS-aided NOMA networks can perform better in V2V wireless communication and can be suitable in B5G-V2V radio networks.

The Internet of Things (IoT) is expanding rapidly as more consumer devices become connected. This expansion presents significant security challenges for Software-Defined Networking (SDN), where static defense mechanisms are often insufficient. The proposed work presents a novel framework that enhances Intrusion Detection Systems (IDS) by combining dynamic resource allocation with game theory and Reinforcement Learning (RL). In this framework, the strategic interactions among the IDS, potential attackers, and the SDN controller are modeled to reflect the complex dynamics of consumer-centric IoT environments. The controller allocates resources in real time, adapting to network demands while maintaining security and performance. RL optimizes these decisions by enabling continuous adaptation to evolving attack patterns. Simulation results validate the effectiveness of the proposed approach, achieving a 98.2% attack detection rate with only 1.8% false positives, along with measurable improvements in throughput and latency compared to baseline methods. The proposed work resolves the key trade-off between security and efficiency, delivering a robust and efficient security framework for IoT-enabled SDN in consumer environments.