Thirumurugan Shanmugam
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Thirumurugan Shanmugam has a total of 12 co-authors. Here are the top co-authors by number of shared publications:
- Arun Kumar Sivaraman
Algoma University
3 shared publications - Nithiyanantham Janakiraman
Department of Electronics and Communication Engineering, K.L.N. College of Engineering, Madurai, India
2 shared publications - Priya Ravindran
K.L.N. College of Engineering,Department of ECE,Tamilnadu,India
2 shared publications - Chirag Chandrashekar
College of Design & Engineering, National University of Singapore (NUS), Singapore, Singapore
1 shared publication - Donipathi Mogili Reddy Prasad
Chemical and Energy Engineering Programme Area, Faculty of Engineering, Universiti Teknologi Brunei, Gadong, Brunei Darussalam
1 shared publication - Govindarajan Lakshmanarao
Department of Engineering, College of Applied Sciences, Sohar, Sultanate of Oman
1 shared publication
Latest Publications
- Open Access
- 25 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 38-42
An IoT Integrated Polymeric Nanocomposite Membranes for Smart Produced Water
- 25 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 38-42
Open AccessAn IoT Integrated Polymeric Nanocomposite Membranes for Smart Produced Water
- 22 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1042-1046
A Novel U-Net Based CNN Algorithm for High-Accuracy Weather Forecasting
Accurate weather forecasting is essential for sectors like agriculture, aviation, and disaster management. However, deep learning algorithms face challenges in prediction accuracy due to issues like vanishing gradients, overfitting, and high computational demands. This research proposes a novel U-Net based architecture utilizing a Convolutional Neural Network (CNN) bottleneck layer to improve weather forecasting. Key features include a skip-connection mechanism, modified weight update rules, Gaussian-mutation operations, and the Adam optimizer for enhanced feature extraction and faster, more accurate predictions. The model was tested using precipitation data from Doppler Weather Radar (DWR) Chennai Radar and weather parameters from European Centre for Medium-Range Weather Forecasts (ECMWF). A dedicated GeoServer facilitates realtime data processing. Experimental results show the proposed algorithm achieves 97.5% accuracy, outperforming CNN and long short-term memory (LSTM) models by 5.84% and 2.41%, respectively.
- 22 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1042-1046
A Novel U-Net Based CNN Algorithm for High-Accuracy Weather Forecasting
Accurate weather forecasting is essential for sectors like agriculture, aviation, and disaster management. However, deep learning algorithms face challenges in prediction accuracy due to issues like vanishing gradients, overfitting, and high computational demands. This research proposes a novel U-Net based architecture utilizing a Convolutional Neural Network (CNN) bottleneck layer to improve weather forecasting. Key features include a skip-connection mechanism, modified weight update rules, Gaussian-mutation operations, and the Adam optimizer for enhanced feature extraction and faster, more accurate predictions. The model was tested using precipitation data from Doppler Weather Radar (DWR) Chennai Radar and weather parameters from European Centre for Medium-Range Weather Forecasts (ECMWF). A dedicated GeoServer facilitates realtime data processing. Experimental results show the proposed algorithm achieves 97.5% accuracy, outperforming CNN and long short-term memory (LSTM) models by 5.84% and 2.41%, respectively.
- 12 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Addressing Security Risks in IoT by Safeguarding Privacy and Networks of Connected Devices
The increase in the use of Internet of Things, otherwise IoT, has impacted numerous industries such as health, facility, and manufacturing. Although connections have become more sophisticated, security and privacy risks have also emerged through connected devices. This research focuses on the security risks posed by IoT devices and seeks to recommend measures to protect these IoT networks. Based on the analysis of the state-of-the-art, the areas that should be further secured are defined, including energy-aware security mechanisms, privacy-preserving protocols, and edge computing vulnerabilities. The results indicate that 78% of IoT devices used in healthcare facilities can be compromised to release private info to unauthorized personnel; 65% of smart building systems are not secure enough when it comes to encryption. Furthermore, the research showed that by implementing artificial intelligence, organizations can decrease security threats by as much as 42 percent in IoT settings. The study then calls for a layered security solution that incorporates energy management, data protection, and the development of a comprehensive security solution for the IoT systems. Further work should be directed to the research of the flexible safety frameworks, which would allow handling new threats in terms of secure functioning of IoT networks.
- 12 November 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Addressing Security Risks in IoT by Safeguarding Privacy and Networks of Connected Devices
The increase in the use of Internet of Things, otherwise IoT, has impacted numerous industries such as health, facility, and manufacturing. Although connections have become more sophisticated, security and privacy risks have also emerged through connected devices. This research focuses on the security risks posed by IoT devices and seeks to recommend measures to protect these IoT networks. Based on the analysis of the state-of-the-art, the areas that should be further secured are defined, including energy-aware security mechanisms, privacy-preserving protocols, and edge computing vulnerabilities. The results indicate that 78% of IoT devices used in healthcare facilities can be compromised to release private info to unauthorized personnel; 65% of smart building systems are not secure enough when it comes to encryption. Furthermore, the research showed that by implementing artificial intelligence, organizations can decrease security threats by as much as 42 percent in IoT settings. The study then calls for a layered security solution that incorporates energy management, data protection, and the development of a comprehensive security solution for the IoT systems. Further work should be directed to the research of the flexible safety frameworks, which would allow handling new threats in terms of secure functioning of IoT networks.
- 3 November 2024
- Published by Springer Nature
- p. 244-255
GAN-Enhanced Multiclass Malware Classification with Deep Convolutional Networks
- 3 November 2024
- Published by Springer Nature
- p. 244-255
GAN-Enhanced Multiclass Malware Classification with Deep Convolutional Networks