ABSTRACT
The continuing advances in sensor technology facilitate data acquisition and data processing with sensor devices. Due to these improvements, the new wireless sensor networks (WSNs) are able to gain more knowledge about the environment and perform more advanced tasks. The majority of the advanced tasks performed by WSNs, such as data fusion, decision making, hypothesis testing, detection, etc., can be interpreted in the form of a stochastic inference problem. An appropriate tool that can be used to build a solution for a general stochastic inference problem is the factor graph (FG). Graphical models such as FGs have been used to model stochastic relationships of the variables involved in an inference problem. In the stochastic inference, the object of interest as well as the sensors’ measurement data is represented by random variables or processes. With the new advanced sensor and wireless technologies, a WSN can work as a distributed processing system for advanced applications.
