Application of neural networks for sensor fusion in a remote sensing satellite

In the attitude control of remote sensing satellites, several sensors are integrated together to ensure a high measurement accuracy. Data fusion is thus required to process the information collected from different sensors. In this paper, a neural network is adopted as a data fusion algorithm. Due to their learning ability, fault tolerance and abstraction ability, neural networks provide a powerful tool for signal processing in the presence of modelling errors, noise and redundant information. The effectiveness of this neural network-based data fusion approach has been demonstrated through a simulation study on the measurement of the attitude and orbit control of a remote sensing satellite used in the China Austronautic Board.