RIDA: A Robust Information-Driven Data Compression Architecture for Irregular Wireless Sensor Networks

In this paper, we propose and evaluate RIDA, a novel informationdriven architecture for distributed data compression in a sensor network, allowing it to conserve energy and bandwidth and potentially enabling high-rate data sampling. The key idea is to determine the data correlation among a group of sensors based on the value of the data itself to significantly improve compression. Hence, this approach moves beyond traditional data compression schemes which rely only on spatial and temporal data correlation. A logical mapping, which assigns indices to nodes based on the data content, enables simple implementation, on nodes, of data transformation without any other information. The logical mapping approach also adapts particularly well to irregular sensor network topologies. We evaluate our architecture with both Discrete Cosine Transform (DCT) and Discrete Wavelet Transform (DWT) on publicly available real-world data sets. Our experiments on both simulation and real data show that 30% of energy and 80-95% of the bandwidth can be saved for typical multi-hop data networks. Moreover, the original data can be retrieved after decompression with a low error of about 3%. Furthermore, we also propose a mechanism to detect and classify missing or faulty nodes, showing accuracy and recall of 95% when half of the nodes in the network are missing or faulty.