Modeling random deployment in wireless sensor networks for infrastructure-less cyber physical systems

Random deployment enables rapid installation of a large number of sensors in locations of difficult access, which is important for many infrastructure-less cyber physical systems (CPS), e.g., for disaster response, battle field and extraterrestrial operations. However, random deployment results in heterogeneous sensor postures with varying antenna orientations, a source of network performance degradation. Although much is known about the effect of antenna orientations in 802.15-based wireless sensor networks (WSNs), most works assume no physical disturbance to antennas. However, heterogeneous sensor postures (e.g., overturned sensors) may cause significant disruption to the functionalities of antennas in addition to the reported variation of received signal strength (RSSI). In this paper, we provide empirical characterization of node-to-node communication in random deployment and present the first radio model that effectively captures the distinctive characteristics. Through simulations using our model, we provide a quantitative analysis of the aggregate effect of random antenna orientations of a large number of sensors with heterogeneous postures. Furthermore, we introduce a prototype sensor package that can be used to reduce the impact of random deployment.