Magellan ProMark 3 and Trimble GeoXT handhelds: A comparison of real-world GPS performance for mapping applications
Summary
This paper compares the performance of two handheld GIS data collection systems: the Trimble® GeoXT™ handheld and the Magellan ProMark 3 handheld. Both the Trimble and Magellan systems are specified to provide submeter accuracy or better for professional GIS mapping applications. The results of the tests conducted show that the GeoXT handheld consistently outperforms the ProMark 3 handheld, in terms of accuracy and flexibility as a GIS data collector.
Trimble Navigation Limited, 10355 Westmoor Drive, Suite #100, Westminster, CO 80021, USA
© 2007, Trimble Navigation Limited. All rights reserved. Trimble, the Globe & Triangle logo, GeoExplorer, GPS Pathfinder, and NetRS are trademarks of Trimble Navigation Limited, registered in the United States and in other countries. EVEREST, GeoXT, NetR5, TerraSync, and TrimPix are trademarks of Trimble Navigation Limited. Microsoft, and Windows Mobile are either registered trademarks or trademarks of Microsoft Corporation in the United States and/or other countries. All other trademarks are the property of their respective owners.
www.trimble.com
October 2007
Introduction
This paper compares the performance of two handheld GIS data collection systems: the Trimble® GeoXT™ handheld from the GeoExplorer® 2005 series and the Magellan ProMark 3 handheld. Both the Trimble and Magellan systems are specified to provide submeter accuracy or better after postprocessing for professional GIS mapping applications. However, as the following test results show, there is a significant difference in the performance of the two systems in terms of: • accuracy of differentially corrected positions in real world environments • integrity and richness of GIS attribute data • suitability for using third-party GPS applications Specifically, testing and analysis showed that: • In optimal environments the GeoXT handheld was at least 23% more accurate than the ProMark 3 handheld with the internal antenna. • In typical urban and rural environments the GeoXT handheld reliably provided submeter accuracy after postprocessing whereas the ProMark 3 handheld could not provide submeter accuracy. • In a real-use scenario, the GeoXT handheld demonstrated reliable and efficient recording of rich GIS data attribution while the ProMark 3 handheld has limited GIS attribution that is susceptible to data entry errors. • In tough canopy environments, the GeoXT handheld provided good accuracy without significant loss of position yield while the ProMark 3 handheld provided 50% poorer accuracy in the same environment. • The ProMark 3 handheld is a proprietary system and so does not support third-party GPS applications.
Testing overview
To determine the performance of the systems for a variety of mapping and GIS applications, testing was performed across a range of environments and using different methodologies. The tests can be broadly divided into the following categories: 1. Open static tests designed to benchmark the systems in optimal GPS conditions. 2. Realistic urban environment tests designed to measure the accuracy performance of the systems in the kinds of conditions encountered daily in utility, engineering and construction, government, local government and mobile GIS applications. 3. Real-use land inspection scenario designed to assess the GIS attribution functionality, ease of use and accuracy of each system in a typical application. 4. Canopy tests to determine the performance of the systems in tough GPS environments, such as those encountered in natural resource and forest management applications.
Testing accuracy
The actual accuracy of positions can be precisely measured only against a known surveyed location or “truth point”. All testing conducted in this paper was performed over truth points surveyed with either a Trimble 5800 GPS receiver or a Trimble S6 Total Station. All rover data presented in this paper has been differentially corrected against data logged by a Trimble NetRS® or NetR5™ reference station. Baseline length varies and is detailed in each test section.
www.trimble.com
2
All accuracy values presented in this paper are Horizontal Root Mean Square (HRMS) values. The HRMS value represents the horizontal distance from truth within which at least 63% of the recorded positions fall. HRMS is the statistical measure used by both Trimble and Magellan when specifying accuracy. Therefore, all accuracies reported in this paper can be easily compared to the accuracy specifications published by each manufacturer: • Magellan specifies submeter real-time DGPS but does not specify postprocessed mapping accuracies for the ProMark 3 handheld in the product datasheet. • Trimble specifies submeter real-time and code postprocessed DGPS accuracy for the GeoXT handheld.
Test 1: Open static tests
The purpose of the open static tests was to benchmark the systems in optimal GPS conditions.
Open static test methodology
The open static test site is an open field with no obstructions over 10 degrees above the horizon in any direction (see Figure 1).
Test Systems
Table 1 shows the system components that were used in the tests described in this paper.
Table 1. System components
Component Handheld Field software Office processing software External antenna
Magellan system Magellan ProMark 3 handheld MobileMapping software MobileMapper Office software Magellan Survey L1 GPS antenna
Trimble system Trimble GeoXT handheld TerraSync™ software GPS Pathfinder® Office software Trimble Hurricane L1 antenna
Figure 1. Open static test set-up
The truth locations are survey pegs approximately two meters apart in the middle of the field. Testing was conducted with both internal and external antennas to determine the best possible performance for each receiver by antenna type. For the internal antenna testing, the handhelds were mounted on range poles using custom brackets that position the center of the internal antenna over the center of the range pole. For the external antenna testing, each antenna was mounted directly on the range pole screw threads. Each range pole was set up over a truth location, and the antenna heights were measured and entered in the appropriate fields of the field software. Forty files, twenty containing a single occupation of two minutes duration, and twenty containing a single occupation of thirty seconds duration were recorded concurrently
For details of software and operating system versions used in the tests, see Appendix I – Test versions.
www.trimble.com
3
on each handheld using the field software. During the test, the receivers remained tracking without interruption while the data was recorded in discrete 30-second and 2-minute blocks. In the office, both sets of data were postprocessed against a number of reference stations over increasing
baselines.
Open static test results
The accuracy results of the processed data are summarized in Figures 2 and 3.
GeoXT and ProMark 3 open static baseline degradation by antenna (30 second occupations)
1
0.8 HRMS Accuracy (m)
0.6
ProMark 3 internal antenna ProMark 3 external antenna GeoXT internal antenna
0.4
GeoXT external antenna
0.2
0 0 100 200 300 400 500 Baseline (km)
Figure 2. Thirty-second occupations for open static test
GeoXT and ProMark 3 open static baseline degradation by antenna (2 minute occupations)
1
0.8 HRMS Accuracy (m)
0.6
ProMark 3 internal antenna ProMark 3 external antenna GeoXT internal antenna
0.4
GeoXT external antenna
0.2
0 0 100 200 300 400 500 Baseline (km)
Figure 3. Two-minute occupations for open static test
www.trimble.com
4
In the 30-second occupation open static test, the GeoXT handheld ranged from 23%-63% more accurate than the ProMark 3 handheld for any given baseline/antenna combination. In this test the internal and external antenna results were similar for a given receiver, however once the occupation was extended to two minutes there is a distinct difference in the performance of the ProMark 3 handheld when used with an external antenna. At this longer occupation the GeoXT handheld’s internal antenna was still 23%-60% more accurate than the ProMark 3 handheld’s internal antenna. However, when used with an external antenna, the GeoXT handheld is 9-27% more accurate than the ProMark 3 handheld at baselines of 88 kilometers and greater. At the zero baseline in this test, the GeoXT handheld with an external antenna is 20% less accurate than the ProMark 3 with an external antenna (21cm compared to 17 cm). The conclusion is that when using either short occupations or the internal antenna the GeoXT handheld outperforms the ProMark 3 handheld in open static environments. Adding an external antenna and using a longer occupation time of two minutes brings the performance of the ProMark 3 handheld to a comparable level with the GeoXT handheld. While the GeoXT handheld mostly outperformed the ProMark 3 handheld in this test, both systems proved to be reliably submeter in an optimal open static environment.
Test 2: Urban environment tests
The purpose of the urban environment test was to assess how the GPS performance varies when the equipment is taken out of the laboratory-like environment of the open static test and into a typical user environment.
Urban environment test methodology
The task was to map a circuit of 18 streetlights to submeter accuracy. The circuit of streetlights represented a typical environment for many urban and suburban GIS data collection applications (see Figure 4). The streetlights were regularly spaced in a light industrial area containing low-rise buildings, some trees, as well as parked and moving vehicles.
Figure 4. Streetlight 4 of 18. The circuit of 18 streetlights represents a typical GIS data collection environment.
To best match typical use, the internal antenna on each handheld was used. A second test was then completed using an external antenna with each handheld. At each streetlight an averaged point with an occupation time of 30 seconds was recorded concurrently on both units. The test
www.trimble.com
5
circuit of 18 streetlights was mapped four times on two separate days to give a total of 72 mapped points for accuracy analysis. In the office, both sets of data were postprocessed against a number of reference stations over increasing baselines.
Urban environment test results
The results of the accuracy testing in urban environments are shown in Figures 5 and 6.
GeoXT vs. Promark 3 accuracy by baseline street light mapping circuit - 30 second occupations - internal antenna
3.00
2.50
2.00 hrms (meters)
GeoXT
1.50 Promark 3 1.00
0.50
0.00 0 50 100 150 200 250 baseline (kms) 300 350 400 450 500
Figure 5. Urban environment baseline degradation with internal antenna
GeoXT vs. Promark 3 accuracy by baseline street light mapping circuit - 30 second occupations - external antenna
0.80 0.70 0.60 GeoXT hrms (meters) 0.50 0.40 Promark 3 0.30 0.20 0.10 0.00 0 50 100 150 200 250 baseline (kms) 300 350 400 450 500
Figure 6. Urban environment baseline degradation with external antenna
www.trimble.com
6
The Trimble system is consistently submeter, with HRMS values for the internal antenna between 57 and 68 centimeters depending on the baseline. In comparison, the ProMark 3 handheld with the internal antenna is at best giving an HRMS of 1.49 meters when processed on the shortest 18 kilometer baseline. On longer baselines, the accuracy is as poor as 2.42 meters. When used with an external antenna, the ProMark 3 handheld was able to match the performance of the GeoXT handheld’s internal antenna, and achieve submeter accuracy. These test results show that when used with the internal antenna the ProMark 3 handheld is unable to deliver submeter mapping accuracy solutions under typical use conditions, and must be used with an external antenna to achieve submeter accuracy in real environments.
Test 3: Real-use scenario in a rural environment
The purpose of the real-use scenario is to test the real-world performance of the systems in a typical land inspection application.
Real-use in rural environment methodology
The rural test site consists of a small holding consisting of six fields (Figure 7).
Land Inspection Test Site
A5 A6
A3
A2
A4
A1
Area by Plot Plot Name Area (sqm) A1 A2 A3 A4 A5 A6 47183 52384 57855 90839 101776 102241
¯
0 50 100 200 Meters
Figure 7. Map of land inspection test site
The boundaries of each field plot were surveyed to centimeter-level accuracy using a Trimble 5800 GPS receiver. The task was to map each plot of
www.trimble.com
7
land, recording attributes similar to those required to apply for governmental subsidies associated with a given land-use. The aim was to accurately map the six plots and record the required attributes as efficiently as possible. The target accuracy was submeter using the internal antennas on the handhelds, as this is the most convenient configuration for traversing rough terrain, opening gates, or negotiating obstacles such as fences. One way to compare the performance of each system would be to use the same field data collection tool on both handhelds, such as the ESRI ArcPad software. However, the ProMark 3 handheld supports GIS data collection only with the Magellan MobileMapping software. It does not support ArcPad or any other GPS datalogging software designed to work with industry standard NMEA data. Therefore, the test was completed using the Magellan MobileMapping software running on the ProMark 3 handheld and the Trimble TerraSync software running on the GeoXT handheld. Two ProMark 3 handhelds and two GeoXT handhelds were used during the test so that each field was mapped twice with each system. To reflect the regular geometry of the plots, only the vertices of the area features were captured using the GPS logging techniques available in each system. The data was postprocessed against a nearby base station 16 kilometers from the test site. The results of the test were first analyzed in terms of data accuracy. Then the results were analyzed in terms of workflow and the different capabilities of each software product.
Real-use in rural environment test results: Accuracy
The accuracy was assessed by measuring the distance from the surveyed boundary corners of each plot to the corners as mapped by the two systems. The six fields had a total of 31 vertices, which gave a total of 62 samples for each system. In addition, the total land area of the surveyed plots was compared to the total land area of the plots as mapped on each system. Table 2 summarizes the accuracy results for each system in this test.
Table 2. Accuracy indicators for real-use test scenario
Indicator HRMS Total error in measured land area Error as a percentage of total land
ProMark 3 269 cm -6167 m2 1.4%
GeoXT 57 cm -565 m2 0.1%
During the test, the ProMark 3 handhelds gave a combined HRMS of over 2 meters while the GeoXT handhelds gave a submeter HRMS of 57 cm. Given the short baseline, the accuracy for the ProMark 3 handheld in this test was actually poorer than the accuracy in the urban environment test, despite the rural environment having a more open view of the sky. The reason for this may be that in the urban test the ProMark 3 handheld was able to average positions for 30 seconds to derive a single point feature. However the MobileMapping software does not allow averaged positions for a single vertex in an area feature, and so the results for the rural test did not benefit from averaging. In contrast, the TerraSync software allows averaged vertices in line and area features, which meant that each vertex recorded using the GeoXT handheld consisted of 30 seconds of GPS data averaged into a single position. This allowed the GeoXT handheld to maintain the same accuracy
www.trimble.com
8
performance in the real-use rural scenario that it had achieved in the urban environment test. The total land area initially measured by the survey receiver was 452,277 square meters (m2). The inaccuracies in the vertices recorded for each plot meant that the total area calculated by the ProMark 3 handheld was underestimated by more than 6,000 m2 (1.4%). The GeoXT handheld also underestimated the total land area, but by less than 600 m2 (0.1%). The ProMark 3 handheld error was therefore 14 times worse than that of the GeoXT handheld.
To be able to collect this data in the field, the MobileMapper Office software was used to create a feature library for use in the MobileMapper software running on the ProMark 3 handheld, and the GPS Pathfinder Office software was used to create a data dictionary for use in the TerraSync software running on the GeoXT handheld. Both office software programs were easy to use and it was a straightforward process to upload the required files to the receivers.
Dictionary size
Real-use in rural environment test results: Workflow
In addition to accurate positioning, the value of a GPS handheld for GIS data collection lies in streamlined workflow and the capture of rich attribute data in a controlled format. These characteristics are discussed first in relation to preparing for data collection in the office and then in relation to collecting data in the field. Preparation for data collection The target database for the data was created using ESRI ArcGIS Desktop software, version 9.1. Table 3 shows the schema for the land inspection database used for this test. It consists of a single feature with nine attributes of various types that are used to describe the feature.
Table 3. Test data schema
The test data schema was very simple—a single feature type and nine attributes. In reality data schema are often much more involved and contain many features and attributes. The exact number of these depends on how the data will be used. This test showed a significant difference in the ability of the two different software products to handle a complex data dictionary. A Magellan feature library can contain a maximum of only 15 features, and each feature a maximum of only 10 attributes. While the Magellan feature library could cope with the design requirements of this test data schema, additional attributes that might be required could easily exceed the limits of the system. For example, a real land inspection application might require additional attributes such as planning designation and rural address codes. In contrast, Trimble data dictionaries have no practical limit on the number of features allowed per dictionary and the number of attributes allowed per feature, which means that the requirements of any data schema are easily met.
Required attributes
Feature Plot
Attribute Name Name Region District Owner Land use type Land use detail Photograph Survey Date Survey Time
Type Text Menu Menu Text Menu Menu File Date Time
Permitted values 100 char 8 choices 10 choices 100 char 4 choices 7 choices Any file Dates Times
To ensure a complete inspection every time, it is helpful to flag all attributes as required, which means they must be entered by the field worker. This ensures key data values are not missed and costly repeat visits are avoided. The Magellan feature library does not allow attributes to be flagged as required, whereas the Trimble data dictionary allows these to be set at the discretion
www.trimble.com
9
of the office manager. In this test, only the Trimble system could guarantee that a full inspection record was completed for each plot.
Date and time attributes
The Magellan feature library editor does not include specific date or time attribute types. These attributes had to be included in the feature library as text fields, with no restrictions available on the format used for entry. This means that different field workers might enter times and dates in different formats, and would be relying on external sources, such as their wristwatch, for the date and time. In contrast, the Trimble system auto-generates both date and time attributes based on GPS time. The user does not need to spend time entering those attributes, and they are recorded in a pre-defined format to ensure consistency between field workers, and with the GIS data schema.
Photograph attributes
The photos were taken with a Nikon P3 CoolPix digital camera and then wirelessly transferred to the GeoXT handheld using TrimPix™ technology. The photo was attached as an attribute to the area feature using the TerraSync software, and ultimately stored as part of the GIS database. For more information on TrimPix technology and the benefits of collecting images as part of GIS data capture, refer to the white paper Adding Digital Photographs to Your GIS. To download this paper from the Trimble website, go to www.trimble.com/geoxt_wp.asp.
Menu attributes
The ProMark 3 handheld does not support files or images as attribute types, and so these could not be collected with the Magellan system. On the GeoXT handheld a file type attribute could be specified and digital photographs of each field included as evidence of land use at the time of inspection. Figure 8 shows one of the photographs taken during the test.
The Magellan feature library editor allows only five menu values to be associated with each attribute. For our test data schema, this was insufficient for the Region, District, or Land Use Detail menu options, and those items had to be entered as text attributes on the ProMark 3 handheld. In reality menu values for attributes in this type of scenario would number in the tens or even hundreds, such as the number of regions in a country or the number of districts in a region. A menu attribute with fewer than six values would rarely be sufficient. Instead the user has to spend time entering the values manually, which increases the risk of different values being entered for the same land use. This results in additional time and effort spent in the office processing and editing the attributes into values that comply with the data schema. However, there is no practical limit to the number of menu values that can be used for a given attribute in the TerraSync software. This enables controlled data collection using the GeoXT handheld and simple office processing after collection.
Text attributes
The Magellan feature library has a limit of 20 characters for a text attribute. Depending on the value of the attribute, this could be insufficient to record an address, region, district, or owner’s
Figure 8. Evidence of current land use for a plot
www.trimble.com
10
name without using abbreviations. This is not an issue with the Trimble system, as the Trimble data dictionary allows up to 100 characters for text attributes. Field workflow The workflow used in the field to complete this test is summarized for each system below:
Trimble workflow
menu items as on the Trimble system. The repeat function helped the ProMark 3 handheld tester to reduce data entry time after the first plot was recorded. However, Repeat had to be selected for every feature on the ProMark 3 handheld rather than just set once and left, as on the GeoXT handheld. When logging the corners of the field, the GeoXT handheld employed averaged vertex logging, which includes a visible position counter as well as an audible logging indicator, so that the user is confident that they are recording data. On the ProMark 3 handheld there is no function to average GPS data into a vertex of an area, so only a single position is required at each corner. This can be achieved by pausing and then resuming logging when positioned over the corner post of the field. However, there is no visible position counter to indicate that a position has been logged, and the audio logging indicator was very difficult to hear in the field with the wind blowing, occasional traffic noise, and so on. The result was a lack of confidence that the position had been recorded, and necessitated putting one’s ear close to the ProMark 3 handheld to hear the logging “beep”.
Map display
1. Start a plot feature with logging automatically paused. 2. Take a photo of the plot (photo transfers wirelessly to the handheld). 3. Enter attributes by selecting options from menus, using the repeat attributes function to populate the feature with default values, and attaching the photo. Date and time attributes are auto-generated. 4. Log GPS data for 30 seconds to create an averaged vertex. 5. Repeat vertex recording at each corner of the field.
Magellan workflow
1. Start a plot feature. 2. Listen for logging audio to indicate that the position is logged and then pause logging. 3. Enter attributes predominantly by text entry, include entering date and time from wristwatch. Repeat attributes function used to populate default values. 4. Resume logging, listen for logging audio to indicate that the position is logged and then pause logging. 5. Repeat logging of position at each corner of the field. The workflows are similar but resulted in a very different user experience. As expected, data entry on the ProMark 3 handheld took more time than on the GeoXT handheld, because the testers could not make use of auto-generated attributes and
One way to confirm that a position has been logged is to check the display of the position in the map window. On three occasions, the map window in the ProMark 3 handheld display indicated that the vertex had been missed by the tester. This meant that the tester had to retrace their steps to the point that was missed and recapture it, resulting in an irregular area that required editing in the office. Once back in the office, it was observed that the points had in fact been collected in the field but had not been drawn correctly in the ProMark 3 handheld map display. The difference between what was shown in the field and what was shown in the office is illustrated in Figure 9.
www.trimble.com
11
Test 4: Canopy test
The purpose of the canopy test was to determine the performance of the systems in tough GPS environments, such as those encountered in natural resource and forest management applications.
Canopy test methodology
The canopy test site consisted of four surveyed points located within a mature pine forest. Each control point was occupied by both receivers (with a minimum antenna spacing of 50 cm) for two minutes, with any data recorded for postprocessing (see Figure 10).
Figure 9. ProMark 3 field map display showing missed positions compared to office map display
Figure 9 shows the “bow-tie” shaped features that appear on the map on the handheld when positions are shown as “missed” and are then recorded out of sequence. To the left of the handheld is the same map downloaded into the MobileMapper Office software. The “bow-tie” features are now triangle-shaped, which occurs when the same position is recorded twice in an area feature. This anomaly was resolved in the office by deleting the additional point logged, but the discrepancy shown in the handheld map had resulted in the ProMark 3 handheld testers unnecessarily walking an additional 1.5 kilometers to re-record points, with a corresponding loss of productivity. The better logging techniques and map display capabilities of the Trimble system meant that the user of the GeoXT handheld could confidently record submeter vertices for each plot. In contrast, the issues with the ProMark 3 handheld’s map display and the lack of logging feedback resulted in less confident data collection and unnecessarily forced testers to return to points and re-record them.
Figure 10. Canopy test antenna setup
www.trimble.com
12
The circuit of four control points was traversed a total of 10 times over two separate days to give a total of 40 occupations for the test. The Trimble system consisted of a GeoXT handheld connected to a Hurricane external antenna. In an effort to track as many positions as possible in this challenging GPS environment, the data was recorded with open GPS quality settings of maximum PDOP of 99, minimum elevation of 0 degrees, and minimum SNR of 12 dBHz. GPS quality filters were applied later in the GPS Pathfinder Office software to impose more stringent quality control levels, and the results were compared. The Magellan system consisted of a ProMark 3 handheld connected to a Survey L1 GPS external antenna. The MobileMapping software does not provide minimum GPS quality settings.
accuracy. These results indicate that the receiver has relatively open masks and a lower level of multipath mitigation compared to the GeoXT handheld, which uses EVEREST™ multipath rejection technology for improved accuracy under canopy. The ProMark 3 handheld was also observed during the test to output two-dimensional (2D) positions which are subject to large errors when the elevation is incorrect. Because this test was designed to show performance in typical natural resource and forestry management applications, it is useful to compare the performance of the two receivers against industry guidelines. The British Columbia Standards, Specifications and Guidelines for Resource Surveys using Global Positioning System (GPS) Technology (Release 3.0, 2001) recommends 10m 2dRMS for mapping grade accuracy in forestry applications. The United States National Map Accuracy Standards (1947) dictate a 12.2 m accuracy at 90% confidence for Hawaii and the Continental US. The GPS Data Accuracy Standard for the US Forest Service expresses this as 13.9 m 2dRMS. In our test, the accuracy of the GeoXT handheld at open masks was within these industry standard accuracy requirements with 10.02 m 2dRMS while the ProMark 3 handheld failed to meet these accuracy standards with 15.46m 2dRMS. It achieved this while maintaining a very high level of productivity, successfully recording 95% of all points. By tightening the GPS quality controls to the Trimble default GPS quality settings, the GeoXT handheld further improved accuracy to 3.67 m HRMS, but with a corresponding lower level of productivity. With these settings the GeoXT handheld achieved accuracy that was twice as good as the ProMark 3 handheld, but achieved only half the rate of productivity. Overall, in allowing a range of GPS quality control settings to be used, the GeoXT handheld offers a more flexible system that allows users to balance their accuracy needs with their productivity goals. The ProMark 3 handheld offers a high degree of
Canopy test results
Data was compared after postprocessing to determine the performance of the two systems under canopy. Table 4 shows the results after postprocessing the data for the two systems.
Table 4. Accuracy and productivity under canopy
Indicator
ProMark 3
HRMS Productivity
7.73 m 100%
GeoXT (open settings) 5.01 m 95%
GeoXT (standard settings) 3.67 m 48%
The GeoXT handheld data was first postprocessed using the same open settings as during data collection, then postprocessed using the default GPS quality settings of maximum PDOP of 6, minimum elevation of 15 degrees, and minimum SNR of 39 dBHz. The MobileMapper Office software has no variable settings for controlling GPS quality, and so the data from the ProMark 3 handheld was postprocessed only once. The ProMark 3 handheld showed 100% productivity but had a corresponding lower level of
www.trimble.com
13
productivity under canopy but at the expense of accuracy.
Conclusion
This paper shows that in these tests the GeoXT handheld comprehensively outperformed the ProMark 3 handheld. The Trimble GeoXT handheld consistently met the manufacturer’s accuracy specifications, reliably providing submeter data not only in optimal open sky conditions, but also in real world urban and rural environments. In optimal environment tests the GeoXT handheld was at least 23% more accurate than the ProMark 3 handheld when using the internal antenna, and at least 9% more accurate than the ProMark 3 handheld when using an external antenna with longer occupation times and baselines. The GeoXT handheld showed reliable submeter accuracy in both urban and rural test environments, while the ProMark 3 handheld was unable to achieve submeter performance in those tests. The GeoXT handheld performed well under canopy, delivering accuracy that meets industry expectations with a high level of productivity (95% yield of corrected positions). Despite demonstrating high productivity under canopy (100% yield of corrected positions), the ProMark 3 handheld failed to deliver acceptable accuracy and had no means of improving accuracy. In the real-use scenario, the GeoXT handheld reliably gave both audio and visual feedback to indicate that positions were being successfully recorded. Map display issues and a lack of visual logging status on the ProMark 3 handheld forced testers to retrace their steps and unnecessarily repeat datalogging. Capturing attribute information was simpler and more reliable using the GeoXT handheld, as the Magellan feature library had a number of limitations that allowed manual data entry errors to be introduced during GIS data attribution. In addition, the GeoXT handheld can be used with
www.trimble.com
14
TrimPix technology to add digital images to the GIS, whereas the ProMark 3 handheld does not support images as attributes. The size limitations of the Magellan feature library may be unsuitable for some applications. In summary, these results demonstrate that the ProMark 3 handheld has simple but limited mapping functionality that may not meet the needs of many GIS field applications. It also did not meet submeter accuracy requirements when used with the internal antenna in real user environments. The results of these tests show that the GeoXT handheld, from the Trimble GeoExplorer 2005 series, is still the most accurate, productive, and easy to use submeter GPS handheld available.
Appendix I – System version information
The tests described in this paper were conducted using the following components:
Magellan system
• ProMark 3 handheld running: − − − ProMark 3 operating system version 1.57 ProMark 3 GPS firmware version P007 MobileMapper receiver software version 1.21
• MobileMapper Office software version 3.40
Trimble system
• GeoExplorer 2005 series GeoXT handheld running: − − − Microsoft® Windows Mobile® 5.0 software GeoExplorer 2005 series operating system version 5.1.14 Trimble GPS firmware 1.81
• TerraSync software version 3.00 • GPS Pathfinder Office software version 4.00
References
British Columbia Standards, Specifications and Guidelines For Resource Surveys Using Global Positioning System (GPS) Technology - Release 3.0 (May 2001), Ministry of Environment, Land and Parks. GPS Data Accuracy Standard (2003) US Forest Service United States National Map Accuracy Standards (1947) US Geological Survey
www.trimble.com
15