9.3.1 Multi-Session Network Processing

INTRODUCTION

The GPS data collected during one session has special properties (essentially, the data is correlated across baselines -- as discussed in section 9.2.1). However, GPS data from different sessions are stochastically uncorrelated -- no double-differences are formed between sessions! This characteristic of session data means that the use of secondary network adjustment procedures to combine different session solutions is largely warranted, without sacrificing accuracy, at the part per million level.

However, before proceeding some comments concerning multi-session GPS networks may be made:

• As in the case of single session network solutions, there are two general classes of multi-session processing techniques:
  • Primary GPS reduction in which all raw carrier phase data, collected over several sessions, is reduced in a one step procedure.
  • Secondary network adjustment using previous single session baseline results as "input".
• A "minimally constrained" (multi-session) network means that only a single datum station is held fixed.
• To combine session solutions, a minimum of one station must be common to two sessions in order to provide the connection between sessions as the network is built up session-by-session (see Figure below). More than the minimum number of connections:
  • increases the redundancies in the survey, and hence improves the accuracy and reliability of the solution, and allows quality control measures to be applied (Figure in section 9.3.2), and
  • results in the secondary network adjustment being sub-optimal vis a vis a primary GPS reduction strategy.
• The manner in which each session is linked to the previous session (and hence back to the original datum station), and to the next session, is an important network design consideration.
• In order to obtain realistic estimates of the coordinate precisions (in fact the entire VCV matrix) it is often necessary to empirically re-scale the stochastic information provided by the primary GPS phase reduction.
• The combination of single session or multi-session solutions (either from primary phase reduction procedures, or secondary network adjustment), into larger campaign solutions will be facilitated through the adoption of a standard solution output format such as SINEX.
• The datum for the resulting multi-session network solution can be further modified, through the determination of the appropriate transformation parameters, so that it may be compatible with that of a local geodetic datum. The network may then be constrained to fit local geodetic control.

The last point is discussed further in section 12.1.1. The other issues raised above will be discussed in this chapter. However it must be emphasised that it is harder to obtain the "best" solution in a one-off adjustment as the number of stations in the network grows. Often different strategies, involving combinations of single baseline and session (with and without redundant baselines) solutions, must be used in an iterative process before arriving at the optimum solution. As with any geodetic network adjustment, coaxing the best results from the available data is as much a "black art" as the rigid application of rules.

A multi-session assembly with no redundancy.

Back to Chapter 9 Contents / Next Topic / Previous Topic

© Chris Rizos, SNAP-UNSW, 1999