In early days of GPS ( early 90s, say) when a full satelite constellation had not matured DOP was more of an issue than today. GPS coverage is lacking in the polar regions (as semi synchronous orbits are only inclinded 55 deg) so DOP will worsen with extreme latitudes with virtually no coverage at the poles. GLONAS which uses more inclined orbits provides much better coverage than GPS in the polar regions something of interest to the Soviets.
Yes, as the satelites move in their orbits DOP geometry changes at a given location on earth with time. Some places may suffer bad DOP while others enjoy good DOP. Should for any reason satelites fail or be shut down for maintenace or have health flags set by ground control, less SVs will be available potentially creating DOP issues in certain places and times. SVs can also be moved reducing coverage at some places while improving in others ( for tactical purposes say remembering GPS is a military system)
Overdetermined solutions are always the aim in GPS (a minimum of 4 of sats will provide position and corrected GPS time - 3 satelites position only) but accuracy and integruty improve as more satelites are added to the solution (the so called all-in-view recievers process all visible satelites - involving 12 or more processing channels) either using standard linear algebra techniques or with a Kalman filter that models user motion. In fact Kalman filters help any GPS rx coast thru periods of bad DOP or loss of signals as can happen when a plane banks or a road vehicle moves into built up or wooded areas that eclipse some of the satelites - the urban canyon effect.
Not surprisingly there is a push , as I write this, to increase the number of active space vehicles for DOP and integrity reasons and of course to have quick deploy orbitting spares.
Last edited by b377; 3rd March 2009 at 09:25.