Distributed Detection and Estimation in Wireless Sensor Networks

Wireless sensor networks (WSN) are receiving a lot of attention from both the theoretical and application sides, in view of the many applications spanning from environmental monitoring, as a tool to control physical parameters such as temperature, vibration, pressure, or pollutant concentration, to the monitoring of civil infrastructures, such as roads, bridges, buildings, etc. [1]. Some new areas of applications are emerging rapidly and have great potentials. A field that is gaining more and more interest is the use of WSN’s as a support for smart grids. In such a case, a WSN is useful to: i) monitor and predict energy production from renewable sources of energy such as wind or solar energy, ii) monitor energy consumption; iii) detect anomalies in the network. A further area of increasing interest is vehicular sensor networks. In such a case, the vehicles are nodes of an ad hoc network. The sensors onboard the vehicle can measure speed and position of the vehicle and forward this information to nearby vehicles or to the road side units (RSU). This information enables the construction of dynamic spatial traffic maps, which can be exploited to reroute traffic in case of accidents or to minimize energy consumption. A relatively recent and interesting application of WSNs is cognitive radio (CR). In such a case, opportunistic (or secondary) users are allowed to access temporally unoccupied spectrum holes, under the constraint of not interfering with licensed (primary) users, and to release the channels as soon as they are requested by licensed users. The basic step enabling this dynamic access is sensing. The problem is that if sensing