A strategy for Geostationary Positioning with GNSS Signals

GNSS signal processing for a receiver in a geostationary environment is more difficult than for a classical receiver on Earth in normal conditions. There are numerous differences between the GNSS signals that an Earth user receives and the signals that a geostationary satellite receives. The specific and main characteristics of the GNSS signal received by a geostationary satellite are the following: high C/No values only for ray tangential to the earth, with C/N0 more frequently between 20dBHz and 24dBHz, very important Doppler values (+/-15kHz), and poor Dilution Of Precision factor (usually higher than 5). To cope with these particular constraints we propose to elaborate a GPS signal processing strategy that has the particularity to be autonomous, that is the receiver does not use aiding data" downloaded from Earth to compute its position. In that case, the complexity of the receiver as well as its cost is lowered. Its integration in the satellite payload is eased too. The geostationary receiver position is computed along the day mainly through an acquisition snapshot process and an orbital filter. As the Doppler range is too large to be processed, as a preliminary step of the strategy, we reduce the Doppler uncertainty by computing the GPS satellite position thanks to almanacs data previously demodulated. This preliminary step is investigated in the first part of the paper. To be able to process enough GPS signals all along the day, with a geometry as good as possible but still limited due to the situation, we implement an acquisition method that is able to cope with C/N0 signal values as low as 20 dBHz. Thus, three techniques have been studied and evaluated: the FFT acquisition method, the Half Bit method and the last one is called the Double Block Zero Padding method. Yet, in parallel to that, we have to ensure that the receiver may be autonomous in that it is able to know enough valid ephemeris data from visible satellites to compute its position thanks to the pseudorange measurements. Indeed, the position of the GPS satellites computed thanks to almanacs data is not enough accurate to compute the geostationary satellite receiver position. It has been shown in [2] that the data demodulation threshold must be lowered down to 25 dBHz to be sure to see at least four GPS satellites with valid ephemeris data at any time during the day. Once we have fixed the demodulation threshold to ensure the autonomy of the receiver, the position is computed through acquisition snapshots. But the position estimate obtained after the acquisition process depends on the accuracy of the peak detection and so, it depends on the sampling rate. The results obtained by using the three above mentioned acquisition schemes do not lead to an accurate position calculation mainly because of the geometry of the acquired GPS satellites. That is why we propose to implement a peak extrapolation technique to improve the accuracy of the peak detection and thus, the accuracy of the estimated pseudorange measurement. Last, in order to improve the position estimate accuracy, we propose to implement an orbital filter to be coupled with the GNSS measurements elaborated at the output of the acquisition process, that takes into account the main forces which drive the movement of a geostationary satellite. The performance of this filter is analyzed in the last part of the paper. The aim of the paper is to present and assess the performance of the global strategy outlined above to process GPS signals with a receiver onboard a geostationary satellite based on the former characteristics. This paper is organized as follows: first, the constraints due to the geostationary orbit are briefly remined, then the strategy is presented in a second part, particularly we stress the means to have an autonomous receiver. A third part is dedicated to the three acquisition techniques and their performances in term of position accuracy and availability are presented. A peak acquisition extrapolation is also detailed. In the last part, a Kalman filter using a simple force model to describe the satellite motion is studied to improve the accuracy of the navigation solution. The computational cost of the strategy is also investigated.