Paul Vincent Craven
All Rights Reserved
Sokkia, out of Overland Park, Kansas donated two high quality GPS's capable of differentiation. GPS systems use a satellite system put into orbit by the United States Department of Defense. By timing the differences in received signals from the satellites, the receiver can calculate its position. The GPS requires three satellites for positioning on a two dimensional plane, and a fourth for calculating the altitude. Sokkia's receivers are capable of receiving six satellites at a time.
Civilian-use GPS's are only accurate to three hundred feet in the horizontal plane, and even less in the vertical plane. Most of this inaccuracy is due to the error that the military automatically puts into the civilian portion of the GPS system. Military GPS's are accurate to approximately ten feet. The reduced civilian accuracy is to help prevent non-military people from using the GPS's for missile guidance. For added protection, the entire civilian portion of the GPS satellites can be turned off, leaving the military GPS system still working. During the time some of the screen shots were taken for this thesis, many satellites were not available for use. This was perhaps because some had moved into position for the Bosnia mission, and/or had their civilian transmissions turned off.
Unfortunately the error introduced into the civilian band is too great for accurate power estimation. A solution to this problem is differentiation. Differentiation uses a stationary base station GPS to cancel out the positional error of roving GPS's. This makes each data reading accurate within ten to thirty feet, according to Sokkia's product specification. Our experience was accuracy within one foot or less. Differentiation isn't quite as simple as it seems. When I asked Sokkia how we could do differentiation in our own code, Sokkia's president, Dr. Joseph Paiva, told me they have spent over a million dollars developing the technology. Another method that can be used for improving accuracy with, or independently from, differentiation is taking repeated readings. Averaging repeated readings of differentiated data can generate even more precise results.
Figure 8: Car power adapter, GPS unit, data logger
Sokkia's hand-held logger is an easy-to-use and durable device that makes it a simple matter to take readings. It runs on either external power or two internal 9-volt batteries. If the internal batteries go bad, two lithium batteries provide memory backup until power can be restored. It apparently runs on an embedded Novell DOS, just like a full-fledged computer. The internal software is updatable using Easily Erasable Programmable Read Only Memory (EEPROMs), so the software on the data logger itself can be updated by a serial link. We had to do this when the logger that was shipped did not include the beta version of their logging software that we needed. Given the chaos of twenty people in three vans and a pickup-trailer, the fact that Sokkia's equipment encountered no problems and required no maintenance work was a pleasant surprise. The cable and cable connectors were high-quality, helping to avoid broken connections.
The first step in taking the needed GPS readings is to set up a base station. This involves first setting up the base antenna. The antenna is on a tripod with a bubble centering device that allows the antenna section of the tripod to be straight up and down. If the antenna were tilted, the position retrieved would be where the antenna was, which would be a few inches off of where the antenna base is. Usually the base antenna would be set up at a known surveyed location. However, since we are only interested in relative accuracy of the data, as opposed to absolute, we did not need to be so careful setting up the base station. We would need to be concerned about relative accuracy if we were laying the GPS data onto a map.
After the base antenna is set up, its coaxial cable is plugged into the GPS receiver. The GPS receiver has a serial cable which then attaches to the notebook computer. Part of the cable also branches off and serves to connect to the GPS's power source. This can be a battery, 12 volt power from a car, or from a wall transformer plugged into an outlet. The base station data is recorded by Sokkia's logging software. The software shows the location of the satellites it is tracking and their relative signals as well. See Figure 9 for a screen shot of the base station recorder.
Attached to the roof of the scout vehicle was a magnetic mount antenna. This antenna does not need the metal disk that the base station antenna has, as the vehicle's metal roof acts as the ground plane for the antenna. The antenna is attached to the hand-held data logger via coaxial cable. The hand-held logger we used held four megabytes of memory, enough for about 250 miles of traveling at 50 miles an hour.
If the scout vehicle looses contact with the satellites it must stop moving. To aid in finding the time when the most satellites will be available, a program is available for the computer that looks at the current almanac transmitted by the satellites. By looking at this almanac, the program can predict when there will or will not be enough satellites in range to take good positional readings. Otherwise it is possible that the scout vehicle may have to wait up to an hour while the satellites are in poor position.
Figure 12 shows predicted satellite geometry over a twenty-four hour period. The geometry is measured using a measurement called DOP. I was never able to find out what DOP stands for, if anything. The lower the DOP, the better the positional readings. Good readings are taken when the DOP is less than four. Looking at Figure 12, we can see that taking readings from 6:00 a.m. to 7:00 a.m., and 2:00 p.m. to 3:00 p.m. will be difficult, if not impossible.
When the scout vehicle returns from logging, the data from the hand-held logger is transferred to the computer holding the base station readings. The computer then takes the two sets of readings and differentiates the scout vehicle's readings, providing the more accurate data set.
After differentiation, the data is converted to a Universal Trans-Mercator
(UTM) format using Sokkia's software. UTM is a universal map projection
whose units are in meters. One can calculate the current UTM zone number
by using the following equation and rounding down (longitude is assumed
to be in degrees):
After differentiation and conversion, the information can be used in the Windows NT™ Power Management System. It can be loaded for power estimation or map display.
A map of the GPS plotted course is available by opening a map window. Map windows can be in an overhead, or in a distance vs. elevation format. Each data reading is shown by a red circle, and the data readings are connected by black lines. One can see the accuracy of the GPS by zooming the path to see the most minute details. Figure 13 shows the detail available through Sokkia's GPS receivers. In this example, we have a car traveling from east to west. Each data reading is taken one second apart, so by looking at the scale and distance between the red dots one can determine how fast the vehicle was traveling. As the car backs up, you can see how the front of the car swings slightly to the left of the original path. Once the car comes to a stop, there is a certain "scattering" of GPS readings where you can see the limitations in accuracy of the GPS. The wider the scattering when the user is stopped, the less accurate the GPS. In this image, the scattering is limited to a couple pixels (a few inches).
The power estimation dialog gives us a close estimation of how much energy we will be using during the rest of the trip given a particular rate of speed. By looking at the weather and altitude, we can get a good indication of how much power we will receive from the solar arrays. Matching the two values gives us the speed at which we should be running. More information on this is given below.