| Distributed sensor systems with the capacity to collect, process, and transmit environmental data have the potential to enable the next revolution in information technology. The rising interest in such sensor systems originates primarily from the low cost of emerging miniature sensing technologies, together with the wide availability of the computing resources necessary to handle complex data. Sensor networks are envisioned to contain legions of wireless nodes. As such, asymptotic regimes where the number of nodes becomes large are important tools in identifying design guidelines for future sensor systems. This work presents interesting applications of large-deviation theory and asymptotic analysis to the design of wireless sensor systems in the context of decentralized detection. Efforts are made to take into consideration the physical components of the communication channels and the structure of the observations available to the sensor nodes. It is found that high node density generally performs well even when observations from adjacent sensors are highly correlated. Furthermore, performance metrics by which sensor node candidates can be compared are established. |
