Izvorni znanstveni članak

https://doi.org/10.17559/TV-20240920001999

Enhanced Data Prediction and Compression in Wireless Sensor Networks Using Bidirectional LSTM with Offset Gaussian Modified Least Mean Square and Renyi-Entropy PCA

Vini Antony Grace N. ; Department of Electronics and Communication Engineering, R.M.D. Engineering College, Kavaraipettai, Chennai, Tamilnadu, India *
Nada Alzaben ; Department of Computer Sciences, College of Computer and Information Sciences, Princess Nourah bint Abdulrahman University, P. O. Box 84428, Riyadh 11671, Saudi Arabia
Faiz Abdullah Alotaibi ; Department of Information Science, College of Humanities and Social Sciences, King Saud University, P. O. Box 28095, Riyadh 11437, Saudi Arabia
Ahmad A. Alzahrani ; Department of Computer Science and Artificial Intelligence, College of Computing, Umm Al-Qura University, Saudi Arabia

* Dopisni autor.

Sažetak

This paper addresses the critical challenge of energy consumption in Wireless Sensor Networks (WSN), which are often deployed in remote areas with limited battery replacement options. To enhance the network's lifetime, a novel approach that combines two key techniques: the Offset Gaussian Modified Least Mean Square (OGMLMS) filter for data prediction and the Renyi-Entropy Principal Component Analysis (PCA) for data compression is proposed. The OGMLMS filter predicts future data values based on historical data, reducing the amount of data that needs to be transmitted. Subsequently, the Renyi-Entropy PCA compresses the predicted data, minimizing the energy required for transmission. The proposed method is implemented using MATLAB and validated with three univariate datasets. The findings highlight substantial enhancements in performance metrics, including Mean Squared Error (MSE), energy efficiency, transmission costs, and compression ratio. These improvements surpass those achieved by traditional algorithms such as Principal Component Analysis (PCA), Least Mean Square (LMS), Auto-Regressive Integrated Moving Average (ARIMA), and hybrid approaches like LMS combined with Renyi PCA, which often struggle with high energy consumption and inadequate data management, leading to reduced network lifetimes. The proposed method effectively integrates data prediction and compression techniques, enhancing energy efficiency and data transmission while maintaining high data quality. The results clearly indicate a notable reduction in energy consumption by up to 14.457%, along with an impressive 99% compression ratio. This enables sensor nodes to operate longer on limited battery resources, making it highly beneficial for remote monitoring applications such as environmental tracking and disaster management. Additionally, the reduced data volume leads to lower transmission costs.

Ključne riječi

data compression; data prediction; wireless sensor networks (WSNs); energy consumption

Hrčak ID:

332855

URI

https://hrcak.srce.hr/332855

Datum izdavanja:

29.6.2025.

Posjeta: 337 *