Subsurface Plume Tracking Using Sparse Wireless Sensor Networks

Abstract

Deployment of sensors for tracking chemical plumes such as those in the subsurface can be quite expensive. A compressed data gathering scheme for chemical plume tracking is presented. With only a fraction (about 25%) of the measurement points required to achieve a given spatial resolution this novel ap-plication of compressed sensing can track the plume within 7% accuracy compared to the case of a full sensor array. The scheme also gathers and re-distributes information of the sensors to the entire network, compress-ing with Discrete Wavelet Transform. The scheme can be used to disseminate global tracking information to sensors as well, with savings in communications by a factor of 5 in average, if such capability is required. The scheme is capable of interpolating randomly missing sensor points with significant accuracy. It also supports data fusion as a simple addition of coefficients requiring no changes to the message length.