Adaptive Compressive Sensing and Data Recovery for Periodical Monitoring Wireless Sensor Networks

Area of Science:

• Wireless Sensor Networks
• Signal Processing
• Data Compression

Background:

• Traditional data gathering in wireless sensor networks (WSNs) relies on raw data collection.
• Compressive Sensing (CS) enables compression-based gathering using data correlations, improving efficiency.
• Existing CS methods struggle with data exhibiting unknown and dynamic sparsity.

Purpose of the Study:

• To develop an adaptive CS data gathering scheme for WSNs.
• To address the challenge of unknown and dynamic signal sparsity in data recovery.
• To enhance network lifetime and energy efficiency in WSNs.

Main Methods:

• An adaptive CS data gathering scheme is proposed.
• The scheme determines current sparsity and sampling rates by re-sampling minimal measurements.
• An adaptive step size variation is integrated with a sparsity-adaptive matching pursuit algorithm for signal recovery.

Main Results:

• The proposed algorithm accurately captures dynamic signal sparsity variations.
• It demonstrates superior recovery performance and convergence speed compared to traditional methods.
• Simulation results show a significantly longer network lifetime than raw data gathering algorithms.

Conclusions:

• The adaptive CS scheme effectively handles dynamic sparsity in WSN data.
• The integrated recovery algorithm improves performance and speed.
• The approach offers substantial energy savings and extended network operational life.