Article: Energy-efficient data reduction and reconstruction schemes to enhance network lifetime in wireless sensor networks Journal: International Journal of Sensor Networks (IJSNET) 2025 Vol.47 No.2 pp.72 - 87 Abstract: Wireless sensor networks (WSNs) are widely used in applications like environmental monitoring, where sensor nodes periodically collect and transmit data. Reducing energy consumption in WSNs is critical and can be achieved by leveraging spatio-temporal correlations to minimise data transmissions. However, lossy transmissions often result in data loss, impacting decision-making accuracy. Balancing reduced transmissions, minimised network overhead, and high prediction accuracy remains a significant challenge. This paper introduces a data prediction model combining one-dimensional convolutional neural networks (1D CNN) and long short-term memory (LSTM) to enhance prediction accuracy. The first stage employs a 1D CNN to extract abstract features from sensor data for one-step prediction. The second stage iteratively uses historical and predicted data for multi-step forecasting. Additionally, an imputation-based data reconstruction scheme integrates MICE, MissForest, and KNNI. Simulation results demonstrate that the proposed methods outperform existing approaches in accuracy metrics (RMSE, R2) and extend network lifetime. Inderscience Publishers - linking academia, business and industry through research

Title: Energy-efficient data reduction and reconstruction schemes to enhance network lifetime in wireless sensor networks

Authors: Sanjoy Mondal; Saurav Ghosh; Sunirmal Khatua

Addresses: A.K. Choudhury School of Information Technology, University of Calcutta, Kolkata, India; Siksha 'O' Anusandhan Deemed to be University, Bhubaneswar, India ' A.K. Choudhury School of Information Technology, University of Calcutta, Kolkata, India ' Department of Computer Science and Engineering, University of Calcutta, Kolkata, India

Abstract: Wireless sensor networks (WSNs) are widely used in applications like environmental monitoring, where sensor nodes periodically collect and transmit data. Reducing energy consumption in WSNs is critical and can be achieved by leveraging spatio-temporal correlations to minimise data transmissions. However, lossy transmissions often result in data loss, impacting decision-making accuracy. Balancing reduced transmissions, minimised network overhead, and high prediction accuracy remains a significant challenge. This paper introduces a data prediction model combining one-dimensional convolutional neural networks (1D CNN) and long short-term memory (LSTM) to enhance prediction accuracy. The first stage employs a 1D CNN to extract abstract features from sensor data for one-step prediction. The second stage iteratively uses historical and predicted data for multi-step forecasting. Additionally, an imputation-based data reconstruction scheme integrates MICE, MissForest, and KNNI. Simulation results demonstrate that the proposed methods outperform existing approaches in accuracy metrics (RMSE, R²) and extend network lifetime.

Keywords: data reduction; missing data; energy efficient; wireless sensor network; WSN; deep learning; data accuracy.

DOI: 10.1504/IJSNET.2025.144564

International Journal of Sensor Networks, 2025 Vol.47 No.2, pp.72 - 87

Received: 13 Aug 2023
Accepted: 10 Nov 2024
Published online: 19 Feb 2025 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article

Title: Energy-efficient data reduction and reconstruction schemes to enhance network lifetime in wireless sensor networks

Keep up-to-date