"Neural networks based approach for data fusion in multi-frequency navigation receivers"

In this paper a novel method to solve the fine synchronization problem in GNSS receivers is presented, The GPS system modernization phase and the Galileo system development will increase signal availability and hence GNSS system-based applications. This variety of uses pervades almost every aspect of GNSS activity and provides the stimulus for its future improvement. For all these causes, in the past few years, there has been a growing interest in the research on the development of techniques and methods for improving the signal reception. Unfortunately, the receiver measurement is usually affected by errors. As already said, in an urban environment the major error source is given by muttipath fading. The proposed method is based on frequency diversity, i.e. the distortions introduced by the channel can be considered different and uncorrelated for sufficiently spaced frequencies. For this reason it is possible to design receivers that, through the usage of multiple frequencies, can improve the reception of SIS, minimizing the distortion effects of the multipath channel. In the proposed system two frequencies in E band, i.e. E5A and E5B, have been considered, and the derived information is fused by using a Neural Network (NN). The NN bases its adaptive fusion on parameters which represent the amount of noise in each of the considered frequencies. Considering the receiver from an higher level, it would be more accurate and efficient if it would be provided by an artificial intelligence that can be developed within the framework of Cognitive Radio devices, the future paradigm for mobile navigation and communication terminals.