A-CS: Adaptive Compressive Data Gathering for Chain-type Wireless Sensor Networks

Inhalt:
Compressive Sensing has been validated as an efficient way for data gathering in chain-type wireless sensor networks (WSNs). Given the sparsity of the sensor nodes’ raw data, each sensor node just needs to send a constant size of data during the multi-hop data gathering, instead of relaying increasing size of data as in traditional data gathering works. However, we notice that the sparsity of raw data in event-involved local region may not be the same as that in the other usual region, which gives us hint that existing compressive sensing based on uniform sparsity of data actually can be improved. In this paper, we propose an adaptive compressive data gathering method A-CS for chain-type wireless sensor networks. Meanwhile, we design an ultra-lightweight localized sparsity variation detection mechanism for determining the event-involved region. Numerous simulation has been conducted, showing the efficiency of the proposed A-CS method.