Brdge Montorng Wth Garmn Handheld Recevers Emly COSSER, Chrs J HILL, Gethn W ROBERTS, Xaoln MENG, Terry MOORE and Alan H DODSON, Unted Kngdom Key words: GPS, handheld GPS recevers, Grngo, Garmn, brdge deormaton montorng SUMMARY GRINGO (GPS RINEX Generator) s a program that has been developed at the IESSG to record the pseudorange and carrer phase output rom 12-channel Garmn handheld GPS recevers n Rnex ormat. It means that data rom Garmn recevers can be post-processed. Garmn recevers are consderably cheaper than survey-grade recevers and so could contrbute to a more aordable montorng system or brdges and other structures. Experments have been carred out to assess the accuracy o postonng wth the output rom the Garmn recevers n both statc and knematc modes. Results are compared drectly to Leca system 500 recevers connected va a spltter to the Garmn recevers. The man challenge s cycle slp detecton and repar, partcularly as Garmn recevers can suer rom hal cycle slps. Ths paper presents the results acheved and accuraces obtanable. General Sesson Emly Cosser, Chrs J Hll, Gethn W Roberts, Xaoln Meng, Terry Moore and Alan H Dodson Brdge Montorng wth Garmn Handheld Recevers 1/12 1 st FIG Internatonal Symposum on Engneerng Surveys or Constructon Works and Structural Engneerng Nottngham, Unted Kngdom, 28 June 1 July 2004
Brdge Montorng Wth Garmn Handheld Recevers Emly COSSER, Chrs J HILL, Gethn W ROBERTS, Xaoln MENG, Terry MOORE and Alan H DODSON, Unted Kngdom 1. INTRODUCTION The Unversty o Nottngham has been awarded a three year grant by the UK s Engneerng and Physcal Scences Research Councl to research the montorng o structures, speccally brdges. One o the research ams was to use sngle requency recevers nstead o dual requency, to develop a more aordable montorng system. As an extenson to the work conducted wth sngle requency recevers, experments have been conducted wth data rom Garmn handheld GPS recevers. Snce the end o SA (Selectve Avalablty) n 2000 the accuraces achevable by GPS recevers n standalone mode have greatly ncreased. Ths had led to great mprovements n postonng qualty achevable by handheld recevers and also reductons n prce. Ths has been coupled wth publc awareness o GPS rsng, so t s now possble to nd handheld GPS recevers on sale n hgh street electroncs shops used by motorsts and outdoor enthusasts. A Leca system 500 survey grade GPS dual requency recever costs 13,500 whch s 20,535 Euros, whle a Leca sngle requency recever reduces the prce to 8,300 (12,625 Euros). Whle the data recorded by these recevers s very relable, they can be too expensve or many montorng applcatons. A Garmn recever on the other hand can be purchased or between 100 and 400 (150 and 600 Euros). The Garmn recevers used or ths experment only cost 189 (283 Euros). Ths paper outlnes the sotware developed at the Unversty o Nottngham, called Grngo, whch extracts raw pseudorange and carrer phase data rom Garmn GPS recevers. Knpos s dual requency knematc GPS processng sotware whch was moded by the authors to process sngle requency data. Ths processng sotware whch has been adapted to process Garmn data s also ntroduced, partcularly ocusng on the cycle slp detecton method used or sngle requency recevers n the context o brdge montorng. Experments carred out to compare the accuraces acheved wth Garmn recevers and Leca survey grade recevers are dscussed. 2. GRINGO GRINGO (GPS RINEX Generator) s a program developed at the IESSG to record the pseudorange and carrer phase output rom a Garmn handheld recever and convert t nto Rnex ormat. Owners o Garmn 12 channel GPS recevers can use the sotware to postprocess the data rom ther recevers. Post-processng s usually only avalable wth expensve survey grade recevers. General Sesson Emly Cosser, Chrs J Hll, Gethn W Roberts, Xaoln Meng, Terry Moore and Alan H Dodson Brdge Montorng wth Garmn Handheld Recevers 2/12 1 st FIG Internatonal Symposum on Engneerng Surveys or Constructon Works and Structural Engneerng Nottngham, Unted Kngdom, 28 June 1 July 2004
Garmn communcatons protocols allow nternal wayponts, tracks and other normaton to be exchanged wth computers or other Garmn recevers. Some o these protocols are well documented but others are not documented by Garmn at all. GRINGO decodes one o the undocumented protocols whch contans raw carrer phase and pseudorange data and logs ths data n Rnex ormat. For more normaton about GRINGO see Hll, et al. (2000) and Hll and Moore (2002). The Garmn recever must be connected to a laptop or data logger, by the seral port. The computer wll then log the raw data n real tme. Fgure 1: GRINGO start up screen Moore, et al. (2002) conduct zero baselne trals wth two Garmn recevers connected va a spltter to the same antenna. Zero baselne trals are useul as they remove a number o error sources assocated wth GPS such as atmospherc eects, satellte ephemers errors and multpath, because the eects o these error sources are the same at both recevers. Ths test was carred out to analyse the ablty o the decodng algorthms n GRINGO, as an ndependent decodng error on one recever would not be ound on the other recever. The zero baselne tral was carred out over 10 mnutes loggng the data at a 1 Hz data rate. The data was processed usng the ambguty xed carrer values n statc mode and a dstance o 0.0001m rom the reerence to the rover recever was recorded. Analysng the raw carrer phase resduals the precson o the raw carrer phase measurement was calculated to be approxmately 0.0014m. It was not possble to carry out a zero baselne tral or the results shown ths paper. The newer generaton o Garmn recevers use only 2 AA cells, and so provde only 3 volts to power an external antenna. Older Garmn recevers used 4 AA cells and so provded more than 5 volts, whch enabled them to power an external survey grade antenna. The Garmn 76 recevers used or the experments n ths project do not have enough power to run an external antenna and so the Leca recevers connected va a spltter ran the antennas. It would be possble to have an external power source runnng the antenna so that the Garmn could record data on ts own. Connectng the Leca recevers to the hot end o the spltter and the Garmn to the cold end allowed the antenna to be powered. By usng a three spltter conguraton, a our recever spltter test was attempted but, snce the spltters were not ampled, the sgnal power was nsucent to enable trackng by the recevers on the cold end o the rst spltter (the conguraton can be seen n Fgure 2. General Sesson Emly Cosser, Chrs J Hll, Gethn W Roberts, Xaoln Meng, Terry Moore and Alan H Dodson Brdge Montorng wth Garmn Handheld Recevers 3/12 1 st FIG Internatonal Symposum on Engneerng Surveys or Constructon Works and Structural Engneerng Nottngham, Unted Kngdom, 28 June 1 July 2004
Antenna Spltter Spltter Hot Cold Hot Cold Spltter Hot Cold Leca GPS recever Fgure 2: The zero baselne conguraton Garmn GPS recever Leca GPS recever Garmn GPS recever 3. GARMIN DATA PROCESSING The carrer phase rom a Garmn recever can slp by hal cycle amounts, whch also means that ambgutes have to be resolved to the nearest hal cycle. It s not known why the recevers slp by hal cycle amounts, but t s possble that they use a sgnal squarng approach to access the carrer. Conventonal sotware packages wll only detect ull cycle slps and so do not cope well wth Garmn data. P4 s statc post-processng sotware that s provded wth GRINGO, whch wll cope wth the hal cycle slps. However, so that knematc data could be processed or the applcaton o structural montorng the n-house sotware Knpos, developed at the Unversty o Nottngham, was moded to cope wth Garmn data. Methods o acceleratng the ambguty search or sngle requency recevers n the context o a brdge envronment were ntroduced nto Knpos processng sotware. For a more detaled explanaton o the sotware development, partcularly the ambguty resoluton algorthms see Cosser, et al. (2004). The sotware was urther moded to resolve nteger ambgutes to the nearest hal cycle and also detect cycle slps at the hal cycle level. 3.1 Cycle Slp Detecton One o the challenges assocated wth processng sngle requency data s the method o cycle slp detecton. In the dual requency verson o Knpos a number o methods were employed to detect cycle slps whch ncluded coarse detecton wth the range resdual method, and more accurate checks wth the onospherc resdual and our observables equatons (Pattnson 2000). Both the onospherc resdual and our observables methods need data rom two requences to be able to detect slps. The range resdual method could be used or sngle requency data, but t can only be used to detect and correct slps o greater than ±4 cycles. A new method o cycle slp detecton needed to be mplemented. In the context o brdge montorng the recevers are not completely knematc. They are movng all the tme, but never by a very large amount. Due to ths act t can be assumed that any large jumps n the carrer phase rom epoch to epoch are lkely to be caused by cycle slps and not recever movement. So, a method based on the trple order tme derence o the carrer phase, δϕ t ), s used to detect cycle slps based on equaton (1) below: ( k General Sesson Emly Cosser, Chrs J Hll, Gethn W Roberts, Xaoln Meng, Terry Moore and Alan H Dodson Brdge Montorng wth Garmn Handheld Recevers 4/12 1 st FIG Internatonal Symposum on Engneerng Surveys or Constructon Works and Structural Engneerng Nottngham, Unted Kngdom, 28 June 1 July 2004
General Sesson Emly Cosser, Chrs J Hll, Gethn W Roberts, Xaoln Meng, Terry Moore and Alan H Dodson Brdge Montorng wth Garmn Handheld Recevers δϕ ( tk) = [ ϕ ( tk) + ϕ ( tk 1)] 3[ ϕ ( tk 1) + ϕ ( tk 1)] + 3[ ϕ ( tk 2) + ϕ ( tk 2)] (1) [ ϕ ( tk 3) + ϕ ( tk 3)] where: ϕ t ) s the carrer phase measurement or satellte on requency at tme t k. ( k ϕ ( t k ) s the carrer phase correcton or satellte on requency at tme t k. Ths carrer correcton s the accumulaton o all the slps on requency that have occurred snce the start o the observaton sesson. I the trple derence, δϕ ( t k ), s larger than a speced threshold then a cycle slp s detected. For Garmn recevers ths threshold s set to be 0.5 cycles and or other recevers t s set to 1 cycle. Through cycle slp smulatons t was shown that ths method could detect all cycle slps at the 1 cycle level, however when t was set to 0.5 cycles a large number o alse cycle slps were ntroduced nto the data. So, the sotware wll now only detect cycle slps o 1 cycle or greater, but or Garmn recevers wll detect them to the nearest hal cycle. I no cycle slp s detected at tme t k then the current cycle slp correcton s set to the prevous one,.e. ϕ ( tk ) = ϕ ( tk 1 ) and no urther calculatons are made or ths satellte at requency. I a cycle slp s detected then the algorthm wll try and repar the cycle slp by usng equaton equaton (2) below: ϕ t ) = ϕ ( t ) + δϕ ( t ) (2) ( k k 1 k where denotes roundng the value to the nearest nteger. A repar s ormed only the varance o the slp value s greater than the sample varance computed rom the thrd order derence. Ths method requres that our epochs o data or each satellte have accumulated beore cycle slp detecton can occur (three epochs or the trple-order tme derence equaton to be ormed and one urther epoch to compute the varance test). For the rst three epochs coarse cycle slp detecton occurs wth the range resdual method. The range resdual RR or requency s calculated usng equaton equaton (3) below (Roberts 1997): ( ρ ( tk ) ρ ( tk 1)) RR = ( ϕ ( tk ) ϕ ( tk 1)) (3) λ where: ρ t ) s the pseudorange measurement or satellte on requency at tme t k. λ ( k s the wavelength or requency For data rom geodetc recevers the range resdual method can detect cycle slps o ±4 cycles or larger due to the nose on the pseudorange. The pseudorange s actually noser or a Garmn recever and so the range resdual method can only be used to detect cycle slps o ±10 cycles or greater. Although the range resdual method can detect cycle slps t s not accurate enough to be able to eectvely correct them. So, durng the rst three epochs a 1 st FIG Internatonal Symposum on Engneerng Surveys or Constructon Works and Structural Engneerng Nottngham, Unted Kngdom, 28 June 1 July 2004 5/12
cycle slp s smply lagged and no attempt at correcton s made. At the ourth epoch a cycle slp can be detected by equaton (1), the trple-order tme derence method, but ths slp wll smply be lagged and not corrected as there s no varance measure to test t aganst. At the th epoch and hgher lagged cycle slps that pass the varance test are corrected. 4. STATIC TRIALS Two statc trals were conducted on the Unversty o Nottngham campus n January 2004. For the rst tral a Leca 510 sngle requency GPS recever and a Garmn handheld GPS recever were connected va a spltter to an AT501 navgaton antenna or a zero baselne tral. For the second tral the same recever conguraton o Leca and Garmn recevers was used at two derence set ups, one or the reerence and one or the rover. The data rom the rst day was processed as a zero baselne tral and also, to nvestgate the raw data qualty, the range resdual varable was examned. On the second day the short baselne was processed rom Leca reerence to Leca rover and rom the Garmn reerence to Garmn rover. Both trals were carred out wth a data rate o 1 Hz (whch s the maxmum possble wth the Garmn recevers). 4.1 Range Resdual The range resdual varable was calculated or the Leca and Garmn data by equaton (3). Ths varable s a good ndcator o the qualty o the pseudorange and carrer phase data rom each recever. The ndvdual pseudorange and carrer phase values were splt nto derent les or the derent satelltes and the range resdual values or each ndvdual satellte were calculated. The range resdual or the Leca data only can be seen n Fgure 3, whle the comparson o the range resduals or the Leca and Garmn recevers can be seen n Fgure 4, both or satellte 16. It can be establshed that the qualty o the raw data or each recever s very derent. Fgure 3and TableTable 1 show that the range resdual or the Leca data s about ±10cm at maxmum, but t s usually around the 3cm mark. The standard devaton or the range resduals s 2.3cm. Ths value s typcal o other results rom the Leca recever. Fgure 4 and Table 1 show that the range resdual or the Garmn data s much worse, reachng 8 metres at maxmum and usually beng around 2-4 metres. The standard devaton n ths case s 1.978m. Leca Range Resdual Range Resduals or the Leca and Garmn recevers 0.15 10 8 Range Resdual (metres) 0.1 0.05 0-0.05-0.1 Range Resdual (metres) 6 4 2 0-2 -4-6 -8-0.15 218000 218500 219000 219500 220000 220500 221000 221500 222000 222500 223000 Tme (GPS Seconds) Leca Fgure 3: The range resdual or the Leca data or satellte 16-10 218000 218500 219000 219500 220000 220500 221000 221500 222000 222500 223000 Tme (GPS Seconds) Garmn Leca Fgure 4: The range resduals or the Garmn and Leca data or satellte 16 General Sesson Emly Cosser, Chrs J Hll, Gethn W Roberts, Xaoln Meng, Terry Moore and Alan H Dodson Brdge Montorng wth Garmn Handheld Recevers 6/12 1 st FIG Internatonal Symposum on Engneerng Surveys or Constructon Works and Structural Engneerng Nottngham, Unted Kngdom, 28 June 1 July 2004
Range Maxmum Mnmum Standard Resdual (m) (m) Devaton (m) Leca 0.096-0.096 0.023 Knpos 7.895-8.651 1.978 Table 1: Summary o results or the range resduals or Leca and Garmn recevers The reason or the huge derences n range resdual values s due to the accuracy and qualty o the pseudorange data. For the Garmn recevers the pseudorange s not very accurate at all. The qualty o the Leca pseudorange data s mproved by pseudorange smoothng whch occurs n the recever tsel. Pseudorange smoothng nvolves usng the more precse carrer phase data to mprove the pseudorange observable. Ths large pseudorange error should not aect the processng o the Garmn data too much as t s the carrer phase that s used manly or postonng solutons. 4.2 Zero Baselne The data rom the Leca and Garmn recevers connected va a spltter to the same antenna was processed on a zero baselne. The Leca recever was used as the reerence whle the Garmn recever was used as the rover. As mentoned prevously a zero baselne tral elmnates many o the error sources assocated wth GPS such as the atmosphere and multpath. Ths test would gve an dea o the accuracy achevable wth the Garmn recevers. Vertcal Error Shown When Garmn and Leca Data are Processed on a Zero Baselne 0.08 0.06 0.04 Error (metres) 0.02 0-0.02-0.04-0.06 218000 218500 219000 219500 220000 220500 221000 221500 222000 222500 223000 Tme (GPS seconds) Fgure 5: The vertcal coordnate error shown when the Leca and Garmn data s processed on a zero baselne The results rom the Garmn and Leca recevers on a zero baselne can be seen Fgure 5 above. In zero baselne trals where two Leca recevers had been used the spread o vertcal coordnates had been around the 2cm level, whch s ntally smlar or ths tral. The unusual thng about Fgure 5 s that there appears to be a slow drt wthn the data. Zero baselne trals rom Leca only recevers do not dsplay ths movement; all the coordnates are evenly spread about the mean value. Ths movement could be attrbuted to the recever clock errors n the Garmn recevers whch are not removed ully by the processng sotware. To nvestgate ths, the clock osets at each epoch were calculated or the Leca and Garmn baselne usng the P4 sotware. The rst General Sesson Emly Cosser, Chrs J Hll, Gethn W Roberts, Xaoln Meng, Terry Moore and Alan H Dodson Brdge Montorng wth Garmn Handheld Recevers 7/12 1 st FIG Internatonal Symposum on Engneerng Surveys or Constructon Works and Structural Engneerng Nottngham, Unted Kngdom, 28 June 1 July 2004
dervatve o the clock oset was calculated and can be seen n Fgure 6 overlayng the postonng soluton. The rst hal o the postonng data has a downward trend whch can also be seen n the clock oset. When the clock oset dervatve starts to latten out the postonng soluton rses. The large jumps n the clock oset dervatve are due to mssng epochs n the Garmn data. It does seem rom the graph that there s a relatonshp between the clock oset dervatve and the postonng soluton. Vertcal Error Shown When Garmn and Leca Data are Processed on a Zero Baselne 0.03 0.0016 0.0015 Error (metres) 0.02 0.01 0-0.01-0.02 0.0014 0.0013 0.0012 0.0011 0.001 0.0009 0.0008 0.0007 Frst Dervatve o theclock oset (ms) -0.03 218000 218500 219000 219500 220000 220500 221000 221500 222000 222500 223000 Tme (GPS seconds) Garmn poston clock oset Fgure 6: The vertcal coordnate error shown when the Leca and Garmn data s processed on a zero baselne overlad wth the rst dervatve o the clock oset 0.0006 4.3 Short Baselne In ths tral two derent short baselnes were processed, one between the reerence Leca recever and rover Leca recever and one between the reerence Garmn recever and the Garmn rover. These two baselnes were dentcal, as the Leca and Garmn recevers were connected va spltters to the same antenna at both ends o the baselne. So, these crcumstances provde a means o drectly comparng the results acheved by the Garmn and Leca recevers. 0.06 Comparson o Vercal Coordnate Error Recorded by the Leca and Garmn Recevers 0.04 Dsplacement (m) 0.02 0-0.02-0.04-0.06 311900 312000 312100 312200 312300 312400 312500 312600 312700 312800 312900 313000 Tme (GPS Seconds) Garmn Leca Fgure 7: The comparson o the vertcal coordnate error recorded by Leca and Garmn recevers over a short baselne Standard Devatons (m) East North Vertcal Garmn 0.0048 0.0139 0.0282 Leca 0.0025 0.0056 0.0135 Rato- Garmn/Leca 1.9472 2.4847 2.0793 Table 2: Summary o the standard devatons or the Leca and Garmn recevers over a short baselne whle the recevers were statc General Sesson Emly Cosser, Chrs J Hll, Gethn W Roberts, Xaoln Meng, Terry Moore and Alan H Dodson Brdge Montorng wth Garmn Handheld Recevers 8/12 1 st FIG Internatonal Symposum on Engneerng Surveys or Constructon Works and Structural Engneerng Nottngham, Unted Kngdom, 28 June 1 July 2004
The vertcal coordnates or the Leca and Garmn recevers can be seen n Fgure 7. It can be seen n ths Fgure that the general pattern o the coordnates s the same or both types o recever, probably due to the multpath characterstcs at the reerence and rover stes. It s obvous however, that the Leca solutons are less nosy than those provded by the Garmn recevers and ths s urther conrmed by the standard devatons that are shown n Table 2. For the short baselne tral the standard devatons o the Leca coordnates are hal that o the Garmn coordnates or the east and vertcal components, wth a bgger derence n avour o the Leca recevers n the north component. Ths s a good result consderng the derence n qualty o the Leca and Garmn raw data. 5. KINEMATIC TRIALS A Leca sngle requency GPS recever and Garmn handheld GPS recever were connected va a spltter to an AT503 choke rng antenna at both the reerence and rover locatons. The reerence locaton was on a known pont on the tower o the IESSG buldng, whle the rover was located on a monument outsde the IESSG buldng, ar enough away so that t had a clear vew o the sky. The rover antenna was located on top o the monument whch had a movable plate. The plate was made to move up and down n the ollowng ways: GPS tme 121352 the plate was made to move downwards approxmately 15cm. GPS tme 121459 the plate was made to move upwards approxmately 15cm. GPS tme 121578 the plate was made to move down and up approxmately 15cm three tmes n successon. GPS tme 121817 the plate was made to move downwards approxmately 15cm. GPS tme 121890 the plate was repeatedly made to move up and down approxmately 2cm or approxmately 100s. GPS tme 122430 the plate was repeatedly made to move up and down approxmately 2cm or approxmately 100s. At all other tmes the plate was statonary. The results can be seen n Fgure 8, whch compares the results recorded by the Garmn reerence and rover, the Leca reerence processed wth the Garmn rover and the Leca reerence and rover. It can be seen rom the graphs the movement o the monument s recorded well by all o the recever combnatons. The movements o 15cm at the begnnng o the observaton sesson are clearly vsble as well as the small dsplacements o 2cm nearer the end o the observaton sesson. The absolute coordnates or the derent recevers, however, are not the same. Both or the Garmn reerence and rover data and also or the Leca reerence and Garmn rover, there s an oset n the absolute coordnates. Ths s caused by errors n the ntal ambguty values, probably because the ambgutes have to be solved to the nearest hal cycle whenever there s Garmn data present. The nterest o the authors s usng Garmn recevers or brdge dynamc deormaton montorng and so the relatve movement o the recevers s o most mportance. I the stuaton were truly dynamc ths oset n coordnates would pose more o a problem. General Sesson Emly Cosser, Chrs J Hll, Gethn W Roberts, Xaoln Meng, Terry Moore and Alan H Dodson Brdge Montorng wth Garmn Handheld Recevers 9/12 1 st FIG Internatonal Symposum on Engneerng Surveys or Constructon Works and Structural Engneerng Nottngham, Unted Kngdom, 28 June 1 July 2004
Vertcal Dsplacement Recorded by the Leca and Garmn Recevers 88.35 88.3 88.25 Dsplacement (m) 88.2 88.15 88.1 88.05 88 87.95 87.9 87.85 121200 121400 121600 121800 122000 122200 122400 122600 122800 123000 123200 Tme (GPS Seconds) Garmn reerence, Garmn rover Leca reerence, Garmn rover Leca reerence, Leca rover Fgure 8: The vertcal dsplacement recorded by the Leca and Garmn recevers or the knematc tral When the recevers are statc the nose n the Garmn data s about twce as bad as the Leca data or ths stuaton also. What s most nterestng s that even wth ths hgh nose value the Garmn recevers stll seem to pck out all the movement o the monument. 6. CONCLUSIONS The Garmn recevers have been tested and compared to Leca geodetc recevers n a number o envronments. The raw range resduals showed a very nosy Garmn pseudorange compared to the phase smoothed Leca pseudorange. A zero baselne tral wth the Leca and Garmn recevers showed a small nose value, but also a drt n the coordnates that s most probably due to uncorrected recever clock errors or the Garmn recevers. On a short baselne the Leca recevers showed results that were twce as good as the Garmn recevers, whch s a good outcome consderng the prce derence or each recever. In a knematc tral the Garmn and Leca recevers showed the same movement, but the absolute coordnates o the Garmn recevers were wrong probably due to ntal ambguty problems caused by the hal cycle values. These ntal trals have been conducted to evaluate the eectveness o usng Garmn recevers to measure deormatons o structures, speccally brdges. It s known that the data rate o 1 Hz s probably too slow to measure all the movement o some structures, or example the Wlord Brdge where a number o trals conducted by the Unversty o Nottngham have taken place (Dodson, et al. 2001). For ths brdge t should be possble to use the Garmn GPS data n conjuncton wth accelerometers to provde a hgher data rate or measurng the dynamc deormatons. For structures or envronmental processes that move slowly, t may be possble to use Garmn recevers to montor ther long term movements. ACKNOWLEDGEMENTS Ths project s unded by the UK s EPRSC n conjuncton wth Craneld Unversty. General Sesson Emly Cosser, Chrs J Hll, Gethn W Roberts, Xaoln Meng, Terry Moore and Alan H Dodson Brdge Montorng wth Garmn Handheld Recevers 10/12 1 st FIG Internatonal Symposum on Engneerng Surveys or Constructon Works and Structural Engneerng Nottngham, Unted Kngdom, 28 June 1 July 2004
REFERENCES Cosser, E., Roberts, G. W., Meng, X. and Dodson, A. H. 2004. "Sngle Frequency GPS or Brdge Delecton Montorng: Progress and Results." 1st FIG Internatonal Symposum on Engneerng Surveys or Constructon Works and Structural Engneerng, 28 June-1 July, Nottngham, England. Dodson, A. H., Meng, X. and Roberts, G. W. 2001. "Adaptve Methods or Multpath Mtgaton and Its Applcatons or Structural Delecton Montorng." Internatonal Symposum on Knematc Systems n Geodesy, Geomatcs and Navgaton (KIS 2001), 5-8 June, Ban, Canada. Hll, C. J. and Moore, T. 2002. "Grngo Sotware; Onlne User Manual." [onlne]. IESSG, Unversty o Nottngham. Avalable at: <http://www.nottngham.ac.uk/essg/grngo > [12 September 2003] Hll, C. J., Moore, T. and Dumvlle, M. 2000. "Carrer Phase Surveyng wth Garmn Handheld GPS Recevers." ION GPS 2000, 19-22 September, Salt Lake Cty, Utah. 178-182. Moore, T., Hll, C. J. and Naper, M. E. 2002. "Rapd Mappng wth Post-Processed Data rom Garmn Handheld Recevers." ION GPS 2002, 24-27 September, Portland, Oregon. 1414-1422. Pattnson, M. 2000. Knpos User Manual, Unversty o Nottngham. Roberts, G. W. 1997. "Real Tme On-The-Fly Knematc GPS," PhD thess, Unversty o Nottngham, Nottngham BIOGRAPHICAL NOTES Ms Emly Cosser s a research student at the Insttute o Engneerng Surveyng and Space Geodesy, Unversty o Nottngham. She graduated n 2001 wth a BSc n Mathematcs rom the Unversty o Nottngham. Snce 2001 she has been workng towards her PhD on brdge montorng wth GPS and other sensors. Dr Chrs Hll s a Senor Research Ocer at the IESSG. He holds a BSc n Cvl Engneerng rom the Unversty o Nottngham and ganed a PhD n Satellte Geodesy, or research on the use o satellte laser rangng or the determnaton o geophyscal parameters. He s a member o the Royal Insttute o Navgaton. Dr Gethn Roberts s a senor lecturer at the Unversty o Nottngham. He s also the charman o the FIG's Workng Group 6.4 "Engneerng Surveys or constructon works and structural engneerng", as well as char o Task orce 6.1.1 "Measurements and Analyss o Cyclc Deormatons and Structural Vbratons". He s also the UK's commsson 6 delegate. Dr Xaoln Meng s a senor research ellow at the Insttute o Engneerng Surveyng and Space Geodesy, the Unversty o Nottngham. He holds a PhD n Hghway, Urban Road and Arport Engneerng rom Tongj Unversty n Shangha, Chna and a PhD n Satellte Geodesy rom the Unversty o Nottngham. Hs research nterests are n engneerng surveyng, satellte geodesy, spatal database development and qualty control, GIS or General Sesson Emly Cosser, Chrs J Hll, Gethn W Roberts, Xaoln Meng, Terry Moore and Alan H Dodson Brdge Montorng wth Garmn Handheld Recevers 11/12 1 st FIG Internatonal Symposum on Engneerng Surveys or Constructon Works and Structural Engneerng Nottngham, Unted Kngdom, 28 June 1 July 2004
Transportaton (GIS-T), Intellgent Transportaton System (ITS), GIS and GPS ntegraton, and GPS, pseudoltes and INS or structural ntegrty montorng. Proessor Terry Moore s Proessor o Satellte Navgaton and Drector o the IESSG at the Unversty o Nottngham. He holds a BSc degree n Cvl Engneerng and a PhD n Space Geodesy, both rom the Unversty o Nottngham. He s a Fellow, and a Member o Councl, o the Royal Insttute o Navgaton and a Member o the (US) Insttute o Navgaton. Proessor Alan Dodson s Proessor o Geodesy at the Unversty o Nottngham, and Head o the School o Cvl Engneerng. He has a BSc n Cvl Engneerng, a PhD n Engneerng Surveyng and a DSc n Engneerng Geodesy. He s a member, and recent Drector, o the IESSG at Nottngham. He s a Fellow o the Insttuton o Cvl Engneers, and the Royal Insttute o Navgaton, a member o the Royal Insttuton o Chartered Surveyors, and an edtor o the Journal o Geodesy. Proessor Dodson was, untl recently, Presdent o Secton 1 (postonng) o the Internatonal Assocaton o Geodesy (IAG) and s a past Dean o the Faculty o Engneerng at the Unversty o Nottngham. CONTACTS Emly Cosser IESSG Unversty o Nottngham Unversty Park Nottngham, NG7 2RD UNITED KINGDOM Tel. + 44 115 951 3933 Fax + 44 115 961 3881 Emal: sxec@nottngham.ac.uk Web ste: http://www.nottngham.ac.uk/~sxec General Sesson Emly Cosser, Chrs J Hll, Gethn W Roberts, Xaoln Meng, Terry Moore and Alan H Dodson Brdge Montorng wth Garmn Handheld Recevers 12/12 1 st FIG Internatonal Symposum on Engneerng Surveys or Constructon Works and Structural Engneerng Nottngham, Unted Kngdom, 28 June 1 July 2004