I rarely do this, but for this column I'm acting as a reporter on two different topics. First, the GPS modernization program has experienced problems that have caused the U.S. Department of Defense to delay the first launch; this update is a follow-up to my March 2005 column. Second, I will take a look at GPS reference station networks (also referred to as virtual or imaginary reference station networks). I'm a geodesist who has taught surveying courses at the university level for a long time, so I'm concerned about the observing procedures used for this development in GPS.
GPS Modernization DelayedAs I reported in my column in March, the eight Block IIR satellites that have not been launched were to be upgraded with a new military M-code on L1 and a second civil signal, L2C, which will be broadcast on the L2 frequency at 1227.6 MHz. After writing the March article I found information to indicate that the first of these satellites would be launched on May 20, 2005. I began checking the U.S. Coast Guard's GPS status daily, and never saw that a new satellite had been launched.
At the beginning of June I checked the website of GPS World magazine and found, under Global View, an article titled "Launch Delay for first Block IIR-M."  This article relates that "a series of glitches has delayed launch of the first modernized Block IIR-M GPS satellite." It goes on to detail the delays: "Originally slated for December 2004, the launch was postponed several times: from December to February 1, then to March 17, then to May 20 (moved up to May 4), and then delayed indefinitely." As this article was going to print, the launch was still unscheduled.
Once the first modernized Block IIR-M GPS satellite is launched, the Air Force will test the new M-code and L2C signals before the transmissions are available for civilian use. For those of us in the surveying community, I would think a minimum of four satellites broadcasting the L2C signal would be necessary for the signal to be truly operational. If we look back at history, the first Block II satellite launched was in February 1989, more than three years after the first shuttle disaster, and four additional Block II satellites were launched in 1989 with five more in 1990. We will have to see what happens.
GPS Reference Station NetworksGPS reference station networks for Real-Time Kinematic (RTK) surveying are beginning to surface everywhere. Articles on these networks and their growing popularity can be read in a variety of professional publications. For example, as noted in this month's Newsline, the French are attempting to establish a network that covers their entire country by the first half of 2006. In June I attended the Arizona Professional Land Surveyors Association annual conference and learned about AZGPS, a company located in Mesa, Ariz., that hosts a network in the Phoenix Valley and offers a subscription service to other firms. eGPS Solutions of Norcorss, Ga. is doing the same (see this month's cover story). Equipment manufacturers are promoting and supporting this technology.
One of my better contacts in the geodetic community is Karen Meckel, a former lead geodesist with the National Geospatial-Intelligence Agency (formerly the National Imagery and Mapping Agency) and current GPS manager for Drexel Barrell & Co. in Boulder, Colo. Karen, like myself, is concerned about a virtual base station that has no repeatability even though the concept is attractive. To quote Karen: "We really are traveling more and more into foggy areas with GPS."
The reason behind her concern is that these networks are able to correct for the ionospheric effects, among other things. At least one manufacturer has the field person occupy the "base" station with a rover, which broadcasts a one-time local position back to the network. The network solves for that "base" position, modeling out the ionospheric effects and creating a tropospheric model for the location of that "base." It then uses these models and that initial position to broadcast "local" corrections to the rover. You don't need a monument or a published position for your "base" because you never return to it (and can never return to it). I heard an interesting comment about this process: "the rover gets tricked into thinking it's getting corrections from a local base station, but it's really getting its info from Big Brother imitating the local base station."
As most RTK surveyors know, nearly all GPS receivers use wide laning to initialize the rover. Wide laning is where a simple function of two phases is used to form the mathematical equivalent of an 86-cm lane (wavelength) . In plain English, this means you don't have the L1 and L2 signals needed to calculate the ionospheric effects on observations, so the distance from the base to the rover has to be relatively short, the same as using an L1-only receiver. These imaginary networks solve this problem.
What about repeatability? At the conference in Arizona I spoke to Wallace Haws, PE, RLS, with HEC Engineering in Mesa, Ariz.; his company uses the network maintained by AZGPS in the Phoenix Valley. The rovers working in their network use Bluetooth technology to transmit to a cell phone, and the cell phone contacts a central server where all data from the individual reference stations is sent.
Haws said there are many established control points in the area, and RTK users occupy these points as checks on their work. He said the repeatability of positions is in the 1 cm range for horizontal and 3 cm range for vertical. This is to be expected; regardless of what method of GPS surveying one uses, the vertical accuracy is about 1/3 that of horizontal accuracy.
I wonder if Dr. Benjamin Remondi, the man who developed the concept of kinematic surveying, ever dreamed it would develop to where it is today.