PCI-MDR: Missing Data Recovery in Wireless Sensor Networks using Partial Canonical Identity Matrix

TL;DR

This paper introduces PCI-MDR, a novel missing data recovery method for wireless sensor networks that leverages partial canonical identity matrices and compressive sensing, outperforming existing techniques on real temperature sensor data.

Contribution

The paper proposes a new method, PCI-MDR, which overcomes limitations of existing approaches by not requiring accurate rank estimation or correlation knowledge.

Findings

PCI-MDR outperforms existing methods in missing data recovery.

The method shows significant improvement on real sensor data.

It effectively handles various missing data scenarios.

Abstract

Data loss in wireless sensor networks (WSNs) is quite prevalent. Since sensor nodes are employed for various critical applications, accurate recovery of missing data is important. Researchers have exploited different characteristics of WSN data, such as low rank, spatial and temporal correlation for missing data recovery. However, the performance of existing methods is dependent on various factors. For instance, correct rank estimation is required for exploiting the low-rank behaviour of WSNs, whereas correlation information among the nodes should be known for exploiting spatial correlation. Further, the amount of missing data should not be massive for exploiting temporal correlation. To overcome the above-mentioned drawbacks, a novel method PCI-MDR has been proposed in this paper. It utilizes compressive sensing with partial canonical identity matrix for the recovery of missing data in WSNs. To validate the proposed method, the results have been obtained on the real data set of temperature sensors from the Intel Lab. The proposed method is observed to perform superior to the existing methods, yielding significant improvement