Cluster based sensor scheduling in a target tracking application with particle filter method
In multisensor applications, management of sensors is neccessary for the classification of data they produce and for the efficient use of sensors as well. One of the most important aspects in sensor management is the sensor scheduling. By scheduling the sensors, serious redictions can be achieved in the cost of bandwith, power and computation. In this thesis, a simple solution for the problem of sensor scheduling in a multi-sensor target tracking application is presented. Proposed method is called sensor grouping. Due to non-linearity and non-gaussianity of the problem itself, proposed solution is presented in the framework of non-linear Bayesian Estimation. For this purpose a detailedtheoretical background of the theory of Bayesian Tracking and Particle Filtering algorithm is given.