SoilNet-TF: A TabNet and DenseNet-121 fused model for soil moisture forecasting and monitoring Veerandra Kumar R, Anbarasan M Agronomy Journal, 2026 Soil moisture is a key parameter for the modeling practices in agriculture, hydrology, and climate, but at the same time, it has to bear the consequences of the drawbacks in conventional methods regarding real‐time accuracy and agility. To make the scenario more friendly, we introduce SoilNet‐TF, an innovative deep‐learning (DL) ecosystem composed of tabular neural network (TabNet) and DenseNet‐121 working together through the attention fusion mechanism that gives the possibility of forecasting and monitoring soil moisture with very high precision and reliability. The dynamic attention of TabNet is applied to the most important soil properties along with sparse attention, whereas DenseNet‐121 allows the reuse of deep features, which leads to the uncovering of the nonobvious interactions among soil parameters. Self‐adaptive pufferfish optimization adjusts feature selection for adaptive real‐time optimization. The model, besides including a bidirectional long short‐term memory for the learning of sequential patterns, also consists of a variational autoencoder for anomaly detection, which aids in accurate soil moisture prediction and provides early warning systems for smart irrigation. The experimental results show a mean absolute error of 0.0129, mean squared error of 0.0013, and R 2 of 0.9728, which means that prediction accuracy has been significantly improved compared to traditional models. The combination of optimization methods and DL in SoilNet‐TF is a very accurate, scalable solution for sustainable agriculture and, subsequently, for water resources management.
An Opamp-Based Coexisting Chaotic Attractors Used in IoT Secure Data Communication M. Anbarasan, Gobalakrishnan Natesan, S. Prakash Journal of Internet Services and Information Security, 2026 In this research paper, an opamp-based simple chaotic circuit is designed exhibiting coexisting chaotic attractors with application in IoT secure data communication is presented. It is interestingly, that this circuit with, only three multipliers and with a few opamps was implemented with OrAD-Pspice, which displays a chaotic attractor. And the coexisting phenomena were also observed by varying initial conditions. This implies it increases the encryption space for IoT secure communication. Finally, an application in IoT for secure data communication, an analog circuit was implemented using the chaotic Masking method. The proposed chaotic masking was verified for single encryption, and double encryption and the original data was decrypted at the receiver successfully using OrAD-Pspice. However, we also prove that when the encrypted signal was decrypted with a different encryption signal other than the transmitted encryption signal, the original signal was not decrypted successfully at the receiver. Also, the proposed coexisting chaotic circuit can be easily manufactured at low cost because of simple circuit implementation compared to nano-based devices such as memristor.
Enhancing Interactive Wireless Communication in Healthcare With Wearable Devices and Smart Healthcare Systems Chaitanya Vasamsetty, Sunil Kumar Alavilli, Bhavya Kadiyala, Rajani Priya Nippatla, Subramanyam Boyapati, M Anbarasan International Journal of Communication Systems, 2026 The analysis for the fifth generation and the communication for the different wireless systems have been analyzed. The specification of the 6G is significantly related to the 5G networks. Some energy consumption‐related problem based on green communication is done to process the data, and the energy consumption is done. Some of the solutions used for the management and the handling of efficient resource are done for providing the technique of the cancellation that can be proposed. The advanced technology used for wireless communication in this technique is the gradient boosting method, the method of hybrid microwave transmission, and so on. Then gradient boosting method is used for monitoring the patient's health in the hospitals using Internet of Things (IoT) devices. The method of hybrid microwave transmission used for analyzing the microwave transmission of the phone calls and the mobile application is evaluated. Effective analysis for data to transfer from one place to another is done. Then 45% of the data is removed based on the redundancy, and the low quality is formed. Furthermore, as seen by its high variance accounted for (VAF) score of 96% after 100 iterations, the suggested technique performs exceptionally well at capturing data variability. It also achieves a quite remarkable ‐squared () score of 0.94, showcasing its strong predictive power and the ability to explain around 94% of the variance of the dependent variable. These figures hence establish the unmatched prediction accuracy exhibited by the proposed method of gradient boosting, explaining data variability and predictability. Trial Registration: We have not harmed any human person with our research data collection, which was gathered from an already published article.
Authorized Block-Mining-Based Intrusion Detection System in Blockchain Enabled IOT Devices Using CDS-KA and BiLeCun-ALSTM Subramanyam Boyapati, Bhavya Kadiyala, Rajani Priya Nippatla, Chaitanya Vasamsetty, M. Anbarasan, Revathi Sundarasekar Journal of Multiscale Modelling, 2026 In the proposed framework, Corollary De-swinging K-Anonymity (CDS-KA) ensures the secure registration and privacy preservation of Internet of Things (IoT) device details, while Bidirectional LeCun Aranda Long Short-Term Memory (BiLeCun-ALSTM) works in tandem within the Intrusion Detection System (IDS) to classify and predict potential attacks. The collaboration between these components enhances both privacy and security, ensuring efficient detection and protection of sensitive data in blockchain-enabled IoT devices. The integration of blockchain technology with IoT devices brings numerous benefits, such as transparency and data integrity. However, it also raises significant privacy concerns. Yet, none of the existing works concentrates on energy-efficient authorized block mining. Hence, this paper proposes an energy-efficient-aware, Authorized Block-Mining-based IDS (ABM-IDS) in blockchain-enabled IoT devices using CDS-KA and BiLeCun-ALSTM. Primarily, the IoT devices are registered using the device details, and then the details are preserved using CDS-KA. At the time of registration, keys and smart contracts are created. The solidity code is used to generate the smart contract. Then, the Merkle tree (MT) is created from the smart contract using GXNOR-BLAKE 512. Also, the solidity functions are split, followed by hash code generation. Then, the generated hash code is updated in the MT. Similarly, the optimal blocks are recognized from the hash code and also verified in the MT. Conversely, the user logs into the network, and then data sensing is done. Thereafter, the data are encrypted and then balanced via Edward Modulo Curve Cryptography (EMCC) and SCC-AZOA, respectively. Now, the balanced data is input to the IDS. In an IDS, the steps such as data collection, pre-processing, feature extraction, feature selection and classification are done. The proposed BiLeCun-ALSTM significantly predicts whether the data is attacked or not. Afterward, the non-attacked data is sent to the destination in a secure manner by verifying the blockchain. Collectively, the proposed framework obtained better security with an accuracy of 98.65%.
An Intelligent Approach for Cloud Infrastructure With Improved Multi-Objective Graywolf Optimization and Resource Allocation for Dynamic Virtual Machine Placement S. Shankar, M. Anbarasan Transactions on Emerging Telecommunications Technologies, 2025 Cloud infrastructure plays a pivotal role in modern computing, yet its optimization and resource allocation often lead to significant delays and power inefficiencies. This research presents an Intelligent Approach for Cloud Infrastructure utilizing Improved multi‐objective gray Wolf Optimization and resource allocation for Dynamic Virtual Machine Placement (ICIMRAD). By mimicking the hierarchical structure and hunting strategies of Gray wolves, the Improved Multi‐objective Gray Wolf Optimization (IMGWO) algorithm, combined with Genetic Algorithms, effectively enhances the accuracy of virtual machine placement and resource allocation. The Fuzzy Group Genetic Algorithm (FGGA) also addresses complex scheduling challenges, facilitating efficient decision‐making across multiple objectives. The dynamic virtual machine system model operates within a Xen environment to monitor power consumption without affecting guest operating systems. Through extensive simulations, the proposed ICIMRAD approach significantly improves metrics such as power consumption, achieving reductions to 0.58 kWh for 50 VMs, and enhances overall system performance compared to traditional optimization methods (e.g., SHOANN, CRASVM, MOOERA). The underlying philosophy emphasizes a powerful synergy between evolutionary strategies and fuzzy logic to drive sustainable and efficient cloud resource management.
A pharmachain IoT with internal attack classification framework using PBFT-MI-RIB-RBF technique in healthcare M Anbarasan, K Ramesh Intelligent Data Analysis, 2025 The pharmaceutical supply chain, which ensures that drugs are accessible to patients in a trusted process, is a complex arrangement in the healthcare industry. For that, a secure pharmachain framework is proposed. Primarily, the users register their details. Then, the details are converted into cipher text and stored in the blockchain. If a user requests an order, the manufacturer receives the request, and the order is handed to the distributor. Labeling is performed through Hypergeometric Distribution Centroid Selection K-Medoids Clustering (HDCS-KMC) to track the drugs. The healthcare Pharmachain architecture uses IoT to control the supply chain and provide safe medication tracking. The framework includes security with a classifier and block mining consensus method, boosts performance with a decision controller, and protects user and medication information with encryption mechanisms. After that, the drugs are assigned to vehicles, where the vehicle ID and Internet of Things (IoT) sensor data are collected and pre-processed. Afterward, the pre-processed data is analyzed in the fog node by utilizing a decision controller. Now, the status ID is generated based on vehicle id and location. The generated status ID is meant for fragmentation, encryption, and block mining processes. If a user requests to view the drug’s status ID, then the user needs to get authentication. The user’s forking behavior and request activities were extracted and given to the classifier present in the block-mining consensus algorithm for authentication purposes. Block mining happens after authentication, thereby providing the status ID. Furthermore, the framework demonstrates an efficaciousness in identifying assaults with a low False Positive Rate (FPR) of 0.022483% and a low False Negative Rate (FNR) of 1.996008%. Additionally, compared to traditional methods, the suggested strategy exhibits good precision (97.869%), recall (97.0039%), accuracy (98%), and F-measure (97.999%).
BRCA-GNN: Graph Neural Network for Breast Cancer Gene Module Discovery M Anbarasan, K. Kokilavani, S Shankar, C. Chandru Vignesh, Stewart Kirubakaran S 2025 3rd International Conference on Sustainable Computing and Smart Systems Icscss 2025, 2025 Breast cancer takes the highest cancer death toll worldwide, with the BRCA1 and BRCA2 gene mutations being some of the most well-studied and critical genetic risk factors. This work proposes a framework called BRCA-GNN for graph neural networks that can identify relevant modules containing the disease genes thorough learning from interaction. It was integrated with high-confidence gene interaction information from STRING and BioGRID to create a biologically relevant graph of approximately 8,000 nodes and 12,000 edges. The graph contains 150 breast tissue samples, consisting of 90 tumor samples and 60 normal samples. Each gene was represented as a node with features corresponding to its level of expression. The edges were known interactions that existed among the genes. This was achieved by a Graph Attention Network (GAT) that captured contextual and topological relationships between genes. It outperformed the standard machine learning baseline, reaching accuracy 87.5, AUC 0.91, precision 0.89, and recall 0.85. Attention mechanisms enabled better predicting and understanding interactions between BRCA1 and BRCA2 and significantly enhanced predictive capabilities. The suggested framework offers an approach for finding functional gene modules that contribute to tumor growth which offers an alternative to traditional classifiers with a structural focus. Also, it assists in later applications such as biomarker discovery and personalized medicine due to its interpretability and modular construction. BRCA-GNN illustrates the applicability of GNNs for genomic data analysis when driven by biologically relevant interaction networks.
IoT-based external attacks aware secure healthcare framework using blockchain and SB-RNN-NVS-FU techniques Ramesh Kuppusamy, Anbarasan Murugesan Technology and Health Care, 2024 BACKGROUND: In recent times, there has been widespread deployment of Internet of Things (IoT) applications, particularly in the healthcare sector, where computations involving user-specific data are carried out on cloud servers. However, the network nodes in IoT healthcare are vulnerable to an increased level of security threats. OBJECTIVE: This paper introduces a secure Electronic Health Record (EHR) framework with a focus on IoT. METHODS: Initially, the IoT sensor nodes are designated as registered patients and undergo initialization. Subsequently, a trust evaluation is conducted, and the clustering of trusted nodes is achieved through the application of Tasmanian Devil Optimization (STD-TDO) utilizing the Student’s T-Distribution. Utilizing the Transposition Cipher-Squared random number generator-based-Elliptic Curve Cryptography (TCS-ECC), the clustered nodes encrypt four types of sensed patient data. The resulting encrypted data undergoes hashing and is subsequently added to the blockchain. This configuration functions as a network, actively monitored to detect any external attacks. To accomplish this, a feature reputation score is calculated for the network’s features. This score is then input into the Swish Beta activated-Recurrent Neural Network (SB-RNN) model to classify potential attacks. The latest transactions on the blockchain are scrutinized using the Neutrosophic Vague Set Fuzzy (NVS-Fu) algorithm to identify any double-spending attacks on non-compromised nodes. Finally, genuine nodes are granted permission to decrypt medical records. RESULTS: In the experimental analysis, the performance of the proposed methods was compared to existing models. The results demonstrated that the suggested approach significantly increased the security level to 98%, reduced attack detection time to 1300 ms, and maximized accuracy to 98%. Furthermore, a comprehensive comparative analysis affirmed the reliability of the proposed model across all metrics. CONCLUSION: The proposed healthcare framework’s efficiency is proved by the experimental evaluation.
