Abstract: | Differential Global Positioning System (DGPS) is an enhancement to Global Positioning System (GPS) that uses a fixed ground-based reference station coordinate to correction user's position accurately by broadcasting the differential correction. There are two different methodologies used to implement a real-time DGPS service, which are 'block shift technique' and 'range correction'. Corrections to the coordinates of the 'block shift technique' are made by comparing known reference station position and instantaneously computed position. 'Range corrections' technique generate corrections to all pseudorange by comparing true to observed range based on the reference station coordinate. The block shift technique is the easiest way to implement, but it is available only if both the remote and base receivers use the same satellite constellation to generate their point solutions. On the other hand, the user of the 'range correction technique' can make any combination of corrected ranges, not just the reference station constellation, so it is popular real-time strategy to most DGPS systems. There are many kinds of GPS receivers and chip-sets which are being currently developed and already released. If one has skilled technique and a high-priced GPS receiver which can provide raw observables in real-time or can read correction message such as RTCM or SBAS, he/she can construct a customized DGPS system whose accuracy is 1-3m. Manufacturers are still producing instruments which operate only in a stand-alone positioning mode, and many chip-sets installed in car navigation kits or mobile phones cannot provide as accurate positions as DGPS results. And geo-tagging which is built into digital camera or hiking device usually does not consider DGPS accuracy. There is no adequate way to improve 10m accuracy of these systems to the DGPS 1m performance if they don't allow users to access to GPS raw measurements, because most DGPS corrections are based on the 'range correction' technique. This paper describes a new 'block shift' DGPS solution, which can correct the already-calculated GPS device positions in position domain. To generate an adequate position-domain correction, we suggests an algorithm to project correction from measurement to position domain. This algorithm is based on a new idea that the relationship between position and measurement correction is deeply coupled with the geometry of satellite constellation, and we can construct an equation to project all the satellites' correction into the rover's position-domain. According to this projection algorithm and system implementation, we can correct the already-calculated stand-alone GPS NMEA result in position domain by adding the correction vector. To verify our new algorithm, DGPS by CP(Correction Projecting), we've constructed a post-processing test using 3 hr CORS rinex data in Korea, and compared the difference in results between the range correction and new block shift method. Because the satellite constellation of the rover (YONS) was different from that of RS (GUMC), the traditional 'block-shift' technique was valid for only 50% of the processing time, and vertical error of it was occasionally up to 20m. However, the 2DRMS error of the DGPS by CP is only horizontally 1.00m and vertically 1.40m which is horizontally 0.06m bigger than that of the range-domain DGPS. We've also demonstrated the availability to apply the DGPS by CP to on-the-shelf receivers. A U-blox LEA-5H model has logged 24hr SBAS-off position in NMEA format, while SBAS-on output has been logged in the other receiver for referencing. Even-though the U-blox receiver has not provided rawdata, DGPS-CP has improved the horizontal 2DRMS accuracy from 2.93m to 1.83m, and vertical 4.65m from to 2.11m. The mitigation performance of DGPS-CP can be found from the result that the mean value of its error is only 0.18m/0.38m(hor/ver), while those of SBAS-off are 0.38m/-2.18m. Referring that the 2DRMS of SBAS error are 2.23m/1.74m (hor/ver) and that mean values are 0.26m/0.04m, we can say that the DGPS-CP performance is pretty comparable to the existing DGPS systems. Finally, we've tested the feasibility to implement the DGPS-CP into a real-time smart phone based on Android OS. A server broadcasts measurement-domain correction with GPS navigation data, our app implemented in Samsung Galaxy Tab projects the correction into position-domain. The result was not as good as that of Ublox receiver, but the vertical 2DRMS error has been reduced from 9m to 1.5m. Currently Korea government is developing GPS correction broadcasting system based DMB, and is scheduled to service it fully in 2012. Our new approach to correct GPS errors in position domain, not in measurement domain, might be a useful way to upgrade our own devices to DGPS available tools without purchasing a new product for the near future system. And when a large-scaled sensor network is constructed in a SBAS-disabled region, GPS raw data is required to apply CORS correction data and to generate a DGPS fix. GPS chipsets with the capability of uploading raw data usually cost more than other chipsets without that option. Equipping every sensing device with the costly GPS chipset to obtain raw data is not a reasonable and economical approach for large-scale deployment of lowcost sensors. By using DGPS-CP, sensors can benefit from DGPS-level accurate positioning without incurring additional hardware cost. |
Published in: |
Proceedings of the 2013 International Technical Meeting of The Institute of Navigation January 29 - 27, 2013 Catamaran Resort Hotel San Diego, California |
Pages: | 486 - 494 |
Cite this article: | Park, Byungwoon, Kim, Y., Yun, H., Kee, C., "DGPS Enhancement to GPS NMEA Output Data: DGPS by Correction Projection to Position-Domain," Proceedings of the 2013 International Technical Meeting of The Institute of Navigation, San Diego, California, January 2013, pp. 486-494. |
Full Paper: |
ION Members/Non-Members: 1 Download Credit
Sign In |