Juan Blanch, Eric Phelts, Kaz Gunning, Todd Walter, Stanford University Lance de Groot, Laura Norman Hexagon Autonomy & Positioning, Canada

View Abstract Sign in for premium content


In GNSS techniques like Precise Point Positioning (PPP), many of the errors affecting the position solution are modelled and estimated over time. In order to use PPP or related techniques for applications that require integrity, we need models of the errors that will lead to upper bounds of the estimated error covariance. In this work, we use the properties of the power spectral density to develop error models that bound autoregressive models with uncertain parameters. As an example, we show how this approach can be applied to the determination of a first order model that bounds the clock and ephemeris errors of GPS for multiple satellites.