Testing, Benchmarking and Noise Estimation Methods for Geodetic GPS Receiver Systems.

Peter Large and Stuart Riley

Abstract:	What is GPS 'receiver noise' ? How can it be reliably measured and quantified ? Characterizing the noise statistics of a particular highperformance GPS receiver design is useful for such purposes as improved stochastic modeling in least squares estimation / Kalman Filtering, and for the purposes of comparison between different receivers. A method commonly employed to estimate receiver system noise is the zero baseline test, where two receivers are connected to the same antenna via a splitter and the zero baseline data processed by double-difference baseline estimation. It is shown that, on a zero baseline, the residual distribution is driven predominantly by the Effective Measurement Bandwidth of the observable. Changes to this parameter lead to significant changes in the residual distribution in a manner which can be accurately predicted from theory. If a comparison is made between identical receiver technology operating in different Effective Measurement Bandwidths, the standard deviation of the zero baseline double difference residuals may differ by an order of magnitude and the conclusion could be drawn that one receiver has significantly 'lower noise' than the other. However, it is shown that if these two receivers are placed on a very short but finite baseline, this order-ofmagnitude difference will no longer be seen in the residuals. In fact, in some cases the differences may be reversed to some degree. This apparent contradiction is explained by the fact that in an integrated GPS receiverantenna system the noise parameters of greatest significance are the multipath error, antenna low noise amplifier noise figure and differential antenna phase center variation. These noise sources are typically many times greater than the thermal noise due to measurement bandwidth, and will dominate the noise statistics in practice. It is also demonstrated that time domain autocorrelation in the observables may lead to zero baseline double difference residual distributions that are distinctly non- Gaussian, due to over-sampling which may take place if the observable is sampled at a frequency higher than the Effective Measurement Bandwidth in which that observable is made.
Published in:	Proceedings of the 14th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GPS 2001) September 11 - 14, 2001 Salt Palace Convention Center Salt Lake City, UT
Pages:	2184 - 2195
Cite this article:	Large, Peter, Riley, Stuart, "Testing, Benchmarking and Noise Estimation Methods for Geodetic GPS Receiver Systems.," Proceedings of the 14th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GPS 2001), Salt Lake City, UT, September 2001, pp. 2184-2195.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In