Recent advancements in federated learning (FL) are paving the way for more energy-efficient communication networks, particularly within the Internet of Things (IoT) sector. A study led by G. Nalinipriya from the Department of Information Technology at Saveetha Engineering College introduces a novel technique called Lifetime Maximization using Optimal Directed Acyclic Graph Federated Learning in IoT Communication Networks (LM-ODAGFL). This research, published in ‘Scientific Reports’, presents a promising approach to enhance the performance of IoT devices while significantly reducing energy consumption.
Federated learning allows multiple devices to collaborate on machine learning tasks without sharing their raw data. This is particularly beneficial in resource-constrained environments, where devices often struggle with processing power and battery life. The LM-ODAGFL technique employs a Directed Acyclic Graph (DAG) model, which efficiently manages the asynchronous nature of device communications in FL networks. By optimizing this model, the technique minimizes additional resource usage, making it an attractive option for commercial applications.
The study highlights the effectiveness of the Archimedes Optimization Algorithm (AOA) in reducing both user energy consumption and the training loss of the FL model. The findings reveal that the LM-ODAGFL model consumes significantly less energy compared to its predecessors, such as SDAGFL and ESDAGFL. For instance, the energy consumption per round on the FMNIST-Clustered dataset ranged from 0.373 to 0.485 kJ, while the Poets dataset showed a consumption range of 16.27 to 20.34 kJ. In contrast, the previous models exhibited energy usage from 0.000 to 1.442 kJ and 0.00 to 63.89 kJ, respectively.
These results indicate a substantial leap in energy efficiency, which is crucial for the growing number of IoT devices deployed across various sectors, including smart homes, healthcare, and industrial automation. The ability to maximize the operational lifetime of these devices not only enhances their performance but also reduces the overall energy footprint, aligning with global sustainability goals.
The commercial implications of this research are significant. Companies investing in IoT solutions can leverage the LM-ODAGFL technique to optimize their devices, leading to lower operational costs and improved battery life. As industries increasingly prioritize energy efficiency, adopting such innovative technologies could provide a competitive edge.
In summary, G. Nalinipriya’s research offers a compelling solution for enhancing energy efficiency in IoT communication networks, with the potential to transform how devices operate and communicate. As the energy sector continues to evolve, the adoption of federated learning techniques like LM-ODAGFL could play a vital role in shaping a more sustainable future.
