A Hybrid Method for Obtaining the Distribution of Report Latency in Wireless Sensor Networks

Abstract

Efficient event reporting is of vital importance in wireless sensor networks (WSNs). This is especially true in critical-time applications, where ensuring an adequate report latency allows a proper reaction from the network to the occurring phenomena. Despite a QoS analysis based on mean report latency is clearly insufficient for critical-time applications, it is oftentimes presented as the main performance indicator. For instance, high percentiles or the whole distribution of the report latency are much better suited. Furthermore, several applications, such as target positioning, require the transmission of a certain number of event packets to accurately characterize occurring events. Building on this, we present a hybrid method for obtaining the report latency distribution of random access (RA) WSN protocols. Here, the distribution of detecting nodes is obtained by simulation and then we perform the mathematical analysis of QoS parameters. Since the proposed method comprises a general structure, it can be easily adapted for analyzing the QoS of a wide range of RA WSN protocols. We use this method for evaluating the report latency, along with the energy consumption of RA protocols in applications that require the transmission of, at least, k event messages. Results show that the combination of higher detection ranges with low congested wireless environments leads to a clear reduction in report latency and event overlooking probability. This leads to an overall performance increase.