Active Wireless Sensing: A Versatile Framework for Information Retrieval in Sensor Networks

Many existing information processing schemes in sensor networks are based on multihop in-network algorithms that require information routing and coordination between nodes and incur excess overhead in latency and energy consumption. In this paper, we develop an alternative and complementary single-hop approach - active wireless sensing (AWS) - in which a wireless information retriever (WIR) queries a select ensemble of nodes to obtain desired information in a rapid and energy-efficient manner. The basic architecture in AWS consists of:(i) a WIR, equipped with an antenna array, interrogates the wireless sensors with wideband space-time waveforms, (ii) the sensors modulate the acquired waveforms with their (possibly encoded) measured data and generate an ensemble response to the WIR's interrogation signal, and (iii) the WIR extracts the sensor data by exploiting the space-time characteristics of the resulting multipath sensing channel. To facilitate analysis, we propose a canonical family of sensing configurations that represent a simple abstraction of spatial correlation in the signal field or the nature of local cooperation in the network. The concept of source-channel matching is introduced in which the spatio-temporal resolution is adapted to the spatial scale of node correlation in the sensing configurations. Signaling strategies and associated receiver structures at the WIR are developed for different source-channel matching configurations. The performance of AWS is analyzed in different configurations both in terms of reliability and capacity of information retrieval.