Measuring the efficiency of the sensing process in a wireless sensor network

Abstract

We apply information-theoretic tools to the design of Wireless Sensor Networks (WSNs). In particular, we propose using mutual information to measure the efficiency of the sensing process in a WSN and use the resulting measure to analyze the efficiency in example WSNs. We show that WSNs have inherent causes of inefficiency due to the manner in which their sensors take measurements. We then use ideas from source coding theorems to suggest more efficient WSN designs. Specifically, we show that the sensing functionality of a WSN with sensors that take noisy measurements can actually be more efficient than a WSN with noiseless sensors.