Experimental assessment of a PPP-based P2-C2 bias estimation

The inception of C2 brought with it new issues to be considered, such as the merger of C2-capable and legacy receivers and the processing of data collected by a C2-capable receiver with satellite clock values generated using a legacy receiver network. Since receiver and satellite hardware delays for C2 measurements may not be necessarily the same of those for P2, a bias between P2 and C2 code measurements should be considered. We refer to this as the P2-C2 bias using the same standard nomenclature used for P1-C1 biases. The knowledge of this bias allows the use of C2 as an observable for positioning, applying satellite clock values computed using P2 as the observable on L2, as in the case of IGS clock products. This paper presents two estimation strategies for obtaining P2-C2 bias. The first strategy takes the mean of the difference between time series of P2 and C2 observations. This strategy relies on receivers tracking simultaneously P2 and C2. The second strategy takes advantage of the residuals of PPP estimation. This PPP-based approach for P2-C2 bias estimation was developed at the University of New Brunswick and has been realized in GAPS, the GPS Analysis and Positioning Software. The P2-C2 values are estimated from a network of C2-capable receivers (a sub-set of the IGS L2C Test Network), and used in the point positioning of a C2-capable receiver, a Trimble R7 receiver, located on the roof of a TU Delft building. The PPP position of the R7 receiver is computed without applying P2-C2 bias, and applying the P2-C2 biases computed from the two approaches. Results show an improvement in the 3D position and their spread of 19% and 3% (1-sigma), respectively, if using the mean P2-C2 values, and of 16% and 8% (1-sigma), respectively, if using the PPP-based P2-C2 bias values.