Sensor data fusion algorithms using square-root information filtering

Abstract

Multi-sensor information fusion is gaining increasing importance in modern technologies. The merit of the information fusion approach is that the fusion at measurement level is quite straightforward. The square-root information filter (SRIF) is studied for possible application to sensor data fusion processes. A decentralised square-root information filter (DSRIF) is proposed. The implementation of the square-root filter should be useful for obtaining numerical accuracy, reliability and stability of fusion algorithms. Results of validation of the DSRIF using simulated data are given.