A genetic algorithm for selection of noisy sensor data in multisensor data fusion

Irrespective of the specifics of a given application, multisensor data fusion problem is mainly composed of three sub-problems: selection, fusion and estimation. Sensor measurements inherently incorporate varying degrees of uncertainty and are, occasionally, spurious and incorrect This, coupled with the practical reality of occasional sensor failure greatly compromises reliability and reduces confidence in sensor measurements. In order to avoid any false inferences, we need data pre-processing methods to make sure that the data to be merged is consistent. Selection of noisy sensor data is a preprocessing of data before merging and is referred to as choosing a representative subset of the sensors that are consistent. In this paper, we use genetic search and optimization approach to develop a genetic algorithm for qualifying the data