United States Department of Agriculture Forest Service
		Technology & Development Program
		September 1996	9671-2341-MTDC
		2200/2300/2400/2600/3400/ 5100/5300/5400/6700/7100

In November 1995, four GPS receivers were tested on the newly established Clackamas GPS test course on the Mt. Hood National Forest. The receivers used were the Trimble Pro XL, GeoExplorer, CENTURION, and the Rockwell Collins PLGR (Precision Lightweight GPS Receiver). The CENTURION and PLGR are PPS (Precise Positioning Service) receivers.

The test course was established by the Pacific Northwest Regional Surveyor after the Forest Service's GPS Steering Group requested him to provide a course in a typical dense West Coast Dougias-fir stand. The course is on gentle terrain in a second-growth Douglas-fir and western hemlock overstory (trees approximately 24 to 40 inches d.b.h.) with a vine maple and red alder understory. It consists of 13 stations, all accurately located within 0.05 meters. The area enclosed by the traverse can be varied from 2.19 to 7.20 acres, depending on the stations chosen.

Reference data for postprocessed differential corrections were obtained from the Portland Community Base Station. It is about 83.8 kilometers (52.13 miles) from the site and was set to log synchronous measurement at 5-second intervals. Differential corrections were applied to the data from the Pro XL and GeoExplorer using two versions of differential correction software, MCORR300 version 3.53 and MCORR400 version 4.03.

Essentially, all receivers were being tested at the same time at different stations on the course with only short periods of double occupancy. The weather during the test was wet with light rain and wet canopy conditions.

CENTURION (PPS-P "Y" code receiver)

Figure 1--CENTURION GPS receiver (left, normally carried in backpack) with data logger (right), and remote antenna.

Average error 6.41 m Maximum error 19.26 m Efficiency 93% Area (point) 7.8% Area (walk) 1.8%

Table 1--Horizontal error and percent error in area for the CENTURION GPS receiver at the Clackamas GPS test course, Oregon, November 1995.

The CENTURION (Figure 1) was connected to the remote dome antenna and data were collected on the MC-V data logger. It was set up with a PDOP mask of 6, SNR mask of 6, elevation mask of 15 degrees, manual 3-D, and a 1-second data logging interval. Approximately 180 position records were collected at each of the 13 stations on the course. The average horizontal position error obtained from this data was 6.41 meters (21.03 feet), with a maximum error of 19.26 meters (63.19 feet) (Table 1). The average collection efficiency under the canopy conditions was 93% (100% = 60 position records per minute). The area was calculated for four different station combinations (loops) that have known areas of 2.19 acres, 4.87 acres, 5.74 acres, and 7.20 acres; the percent error in area calculated was 13.59%, 6.44%, 6.14%, and 4.88%, respectively. The average error in area was 7.8%. However, the stations used to make the smallest loop (13.59% error in area) contained the station with the largest horizontal position error (19.26 meters); none of the other loops did. The largest loop (7.20 acres) was traversed twice, once with data collected in a walk file at the 1-second interval, (1.94% error in area) and once at a 5-second interval (1.71% error in area) for an average of 1.8%. This difference is probably not significant, given the limited data and the PDOP changes with time. Overall, the CENTURION performed well under these canopy conditions.

PLGR (PPS-P "Y" Code Receiver)

Figure 2--Rockwell Precision Lightweight GPS Receiver (PLGR).

CONT to AVG STBY to AVG Average error 7.95 m 8.71 m Maximum error 19.58 m 16.15 m Efficiency -- -- Area (point) 11% -- Area (walk) -- --

Table 2--Horizontal error and percent error in area for the PLGR (Precision Lightweight GPS Receiver) at the Clackamas GPS test course, Oregon, November 1995. The receiver was operated in two modes, continuous to average (CONT to AVG), and standby to average (STBY to AVG).

Two PLGR receivers (Figure 2) were used on the course. One operated in the continuous to average (CONT to AVG) mode of operation, and the other in the standby to average (STBY to AVG) mode. The STBY to AVG mode is recommended for weak signal environments such as under a dense forest canopy. When averaging is selected from the standby mode (STBY to AVG), the receiver stops determining positions and goes through a complete signal search process before determining positions again. This is time consuming, especially under the canopy, and it must be done at each station. When averaging is selected from continuous mode (CONT to AVG), the receiver simply starts averaging with no loss of time. The PLGR operating in the CONT to AVG mode completed the course, averaging approximately 180 position records at each of the 13 stations. The other unit, operating in the STBY to AVG mode, averaged approximately 180 position records at six of the stations before it quit operating. We were not able to find out why it quit operating; later on, it worked fine. When in this mode, the receiver automatically starts the satellite search process using only satellites broadcasting the "Y" code before obtaining a position. This mode is time consuming because it is the same as turning the receiver on from a cold start and it must be done at each station before it can start the averaging process.

For the receiver operating in the CONT to AVG mode, the average horizontal position error was 7.95 meters (26.08 feet), based on the average of approximately 180 position records at all 13 stations, with a maximum error of 19.58 meters (64.24 feet, Table 2). The receiver operating in the STBY to AVG mode produced an average horizontal position error of 8.71 meters (28.58 feet) and a maximum of 16.15 meters (52.99 feet) based on the six stations where it averaged 180 position records. The average for the same six stations from the receiver operating in the CONT to AVG mode was 7.65 meters (25.10 feet). This data would indicate the STBY to AVG mode, for operation under heavy canopy, is not as accurate as the CONT to AVG mode, and is more time consuming.

The data from the receiver operating in the CONT to AVG mode were used to calculate areas for five different station combinations (loops) that have known areas of 2.19, 2.33, 4.87, 5.74, and 7.20 acres; the percent error in the area calculated was 11.17%, 22.79%, 4.44%, 5.74%, and 10.30%, respectively. The overall average error in area was 11%. No attempt was made to collect a walk file with the PLGR due to the time required to manipulate the data for the area calculation.

The operating efficiency of the PLGR could not be determined because in the averaging mode time duration is not recorded.

GeoExplorer (SPS-C/A Code Receiver)

Figure 3--Trimble GeoExplorer GPS receiver.

SNR 4 SNR 6 MCORR300 MCORR400 MCORR300 MCORR400 Average error 7.85 m 7.86 m 4.65 m 4.68 m Maximum error 14.51 m 15.89 m 9.27 m 8.85 m Efficiency 48.3% 48.3% 25.4% 25.3% Area (point) -- -- -- -- Area (walk) -- -- -- --

Table 3--Horizontal error for the Trimble GeoExplorer GPS receiver at the Clackamas GPS test course, Oregon, November 1995. The SNR mask sets the threshold value for the weakest satellite signal the receiver can use when calculating position. The lower the number, the weaker the signal the receiver will track. The MCORR300 and MCORR400 software is used for differential correction.

Two Trimble GeoExplorers (Figure 3) were tested on the course. Both were set up with an elevation mask of 15 degrees, PDOP mask of 6, data logging interval of 1 second, manual 3-D operation. One had an SNR mask set at 4, the other at 6. The SNR mask sets the threshold value for the weakest satellite signal the receiver can use when calculating position. The smaller the number, the weaker the signal that the receiver will track.

Neither receiver was able to collect the planned 180 or more position records at all the stations. This was due to the time required to collect the data because of the canopy attenuation of the signal and the limited time available. The receiver with the SNR set at 4 was able to collect between 120 and 260 position records at 8 of the 13 stations. When this data was differentially corrected, the average horizontal error was 7.85 meters (25.76 feet) for the eight stations with a maximum of 14.51 meters (47.61 feet) (Table 3).

The data were differentially corrected using MCORR300 and MCORR400. However, there was no significant difference in the average horizontal error, 7.85 meters (25.76 feet) for MCORR300 and 7.86 meters (25.79 feet) for MCORR400. The receiver average operating efficiency under the canopy was calculated to be 48.3% (100% = 60 position records per minute). The receiver with the SNR set at 6 was able to collect between 159 and 182 position records at five stations. When this data was differentially corrected, the average horizontal error was 4.65 meters (15.26 feet), with a maximum error of 9.27 meters (30.41 feet). The operating efficiency was 25.4%.

This difference in results between the receivers is as expected. The receiver with the higher SNR setting produced better accuracy because it tracked the stronger signal, while the receiver with the lower SNR setting could track weaker signals that had the potential for noisier signals producing less accurate position data. Because the receiver with the lower SNR setting would track and use the weaker signal, it was more efficient in obtaining positions fixes under the canopy. Areas were not calculated due to the limited data obtained.

Pro XL (SPS-C/A Code Receiver)

Figure 4--Trimble Pro XL GPS receiver with backpack receiver, handheld data logger, and remote antenna.

Wet canopy SNR 4 Dry canopy SNR 6 MCORR300 MCORR400 MCORR300 MCORR400 Average error 5.37 m 5.00 m 2.68 m 1.34 m Maximum error 10.55 m 9.34 m 4.67 m 2.16 m Efficiency 63% 63% 44% 44% Area (point) -- 3.5% -- -- Area (walk) -- -- -- --

Table 4--Horizontal error and percent error in area for the Trimble GeoExplorer GPS receiver at the Clackamas GPS test course, Oregon, November 1995. The SNR mask sets the threshold value for the weakest satellite signal the receiver can use when calculating position. The lower the number, the weaker the signal the receiver will track. The MCORR300 and MCORR400 software is used for differential correction.

The Trimble Pro XL GPS receiver (Figure 4) was set up with an elevation mask angle of 15 degrees, PDOP mask of 6, SNR mask of 4, 1-second data logging interval, and 3-D manual mode of operation. It collected approximately 180 position records at all 13 stations. This data was differentially corrected using both MCORR300 and MCORR400; the average horizontal position error was 5.37 meters (1 7.62 feet) and 5.00 meters (16.41 feet), respectively (Table 4). The maximum error was 10.55 meters (34.61 feet) with MCORR300 and 9.34 meters (30.64 feet) with MCORR400. The operating efficiency under the wet canopy conditions and an SNR of 4 was 63% (100% = 60 position records per minute). Areas were calculated for five different loops that have known areas of 2.19 acres, 2.33 acres, 4.87 acres, 5.74 acres, and 7.20 acres; the percent error in area calculated was 1.6%, 4.25%, 3.13%, 5.74%, and 3.49%, respectively. The average error was 3.5%. No walk files were collected during this test.

For comparison purposes, data collected on the course by the Regional Surveyor during the dry season is shown in Table 4. The Pro XL had been set up the same, except that the SNR was set at 6 and the canopy was dry. The data is from six stations on the course where the average of 118 to 188 position records was obtained. The receiver operating efficiency under dry canopy conditions and the SNR of 6 was 44%. The area of the traverse was not determined because sufficient data were lacking at several stations.

From the data for the Pro XL shown in Table 4, MCORR400 gave better results than MCORR300, whether the canopy was wet or dry. However, there is considerable difference in the average error between dry canopy with receiver SNR mask set at 6 (1.34 meters average error) and wet canopy with SNR set at 4 (5.00 meters average error). This difference could be due to the lower SNR setting and multipath reflection from the wet canopy; both would increase the position error. The lower SNR setting would allow the receiver to track and use a weaker signal with more noise, increasing the potential for position error.

The wet canopy could cause the signal to reflect off a wet surface (multipath) to the receiver rather than line of sight, which would increase the position error. The lower SNR and wet canopy (multipath) would provide the receiver more opportunity to obtain a position fix, increasing operating efficiency, but reducing accuracy compared to a higher SNR and dry canopy. The data would indicate that dry canopy conditions and an SNR mask set at 6 produced the most accurate data.

Results

For both SPS-GPS receivers, the higher SNR mask setting of 6 produced better horizontal position accuracy than an SNR of 4, (Tables 3 and 4). Even though the average horizontal position error of 1.34 meters (4.40 feet) for the Pro XL was under a dry canopy and the GeoExplorer's accuracy of 4.68 meters (15.36 feet) was under a wet canopy, the results were as expected. The higher SNR mask setting means the receiver must track a stronger signal, producing better accuracy. For better accuracy, use an SNR mask setting of 6. The SNR mask setting has the opposite effect on the receiver's operating efficiency under the canopy. Generally, the signal strength under the canopy is weak due to canopy attenuation, which means the receiver with the lower SNR will be more efficient (Tables 3 and 4). Data from the GeoExplorer were corrected with MCORR300 arid MCORR400 with very little difference in horizontal accuracy. The MCORR400 differential correction software was more accurate with the Pro XL (Table 4). This is probably due to the technology used in the receivers. The GeoExplorer uses Gauss-based technology and the Pro XL uses Maxwell-based technology. The reason for the Pro XL's difference in accuracy under the dry canopy, 1.34 meters (4.40 feet), compared to the wet canopy, 5.00 meters (16.41 feet), is probably due to the signal multipath reflection under the wet canopy and the lower SNR mask setting.

The results obtained from this evaluation are based on limited data. Further testing is planned to verify these results.

For further technical information, contact:

Missoula Technology and Development Center
Building 1, Fort Missoula
Missoula, MT 59804-7294
Phone (406) 329-3922
FAX: (406) 329-3719

Hard Copies of this document may be ordered from:

Missoula Technology and Development Center
Building 1, Fort Missoula
Missoula, MT 59804-7294
Phone: (406) 329-3978
Email Document Request