Determining and measuring the true impact of C/A code cross-correlation on tracking—Application to SBAS georanging

This paper will review the necessary conditions for having a code tracking bias due to cross-correlation (XC), focussing on the current C/A code, and show that it will be indeed very unlikely for signals coming from GPS satellites, but it can happen for signals coming from satellites having a low dynamic, such as SBAS or IGSO satellites. Measurement campaigns on SBAS have been made and a surprising result has been found: even though the measured error envelope complies with the theory, the nature of the error is noisy whereas it should have been a bias. The reason for this was explained: this is due to carrier phase jitter on the SBAS ranging signal. The physics of the cross-correltion error will be reviewed. It has many analogies with the multipath effect. The condition for XC errors to happen will be detailed. They are : Doppler collision, data message similitude, and presence of a XC peak. The situation in which XC tracking error happen will be detailed. There are almost impossible for GPS signal due to the satellite high Doppler, but more likely for SBAS satellites with low motion. For WAAS, the message frames of the 2 satellites are most of the time correlated with a correlation index between 0.8 and 1, for EGNOS the messages are sometimes correlated, but most of the timre not, depending on whether the same message type are broadcast at the same time by the satellites. The message symbol have to be received at the same time for the XC error to be maximum. If there is a one symbol shift, that is a 2 ms or a 600km relative code delay, then the received symbols becomes uncorrelated. This explains that XC error can happen only on a very narrow strip, a few hundred km wide only. Additionnaly, within this strip the user must be positioned in a place corresponding to a XC function peak when the Doppler collision happens. A test campaign was made inside this strip. The observed error envelope was found to match the predicted error, but it was noticed an important carrier phase jitter that reached several carrier cycles. This caused the XC error bias to be changed into a random error. This allow to filter out the XC error with carrier smoothing. There is a paradoxal situation : if the carrier phase jitter anomaly was improved, the XC error would become a bias that couln't be filtered with carrier smoothing. If SBAS signal shall be used for ranging, XC error could be taken into account at the receiver level. There are many possibilities, one could be to ignore code measurement in case of Doppler collision, another could be to correct it with an error model.