Comparison of Assisted GPS _AGPS_ Performance using Simulator and

Document Sample
Comparison of Assisted GPS _AGPS_ Performance using Simulator and Powered By Docstoc
					         Comparison of Assisted GPS (AGPS)
     Performance using Simulator and Field Tests
                                                    Sanjeet Singh
                                   Position Location And Navigation (PLAN) Group
           Department of Geomatics Engineering, Schulich School of Engineering, University of Calgary, Canada

BIOGRAPHY                                                       the limitations of the HSGPS receiver which has led to the
                                                                development of AGPS receivers.
Mr. Sanjeet Singh is an M.Sc. candidate in the
Department of Geomatics Engineering, Schulich School            INTRODUCTION
of Engineering, University of Calgary. He was born in
Fiji and obtained a B.Sc. in Electrical Engineering at the      The FCC-E911 mandate and LBS applications for mobile
University of Calgary. He did his internship as a software      users have been driving for accurate positioning solutions,
engineer at Nortel Networks. His current research focuses       which should be able to work anywhere and all the time.
on High Sensitivity and Assisted GPS.                           The FCC-E911 phase II mandate requires an accuracy of
                                                                50 m 67% of the time and an accuracy of 150 m for 95%
ABSTRACT                                                        of the time for handset-based positioning technologies
                                                                (FCC, 2000). Cell phone operators are required to comply
The Federal Communications Commission (FCC) E-911               with phase II by the end of last year. There has been an
mandate, Location-Based Services (LBS), as well as              increase in demand for LBS which should be able to work
personal and vehicular navigation applications are driving      anywhere all the time, applications includes mobile
the need for navigation capability in degraded signal           gaming, child-locator used by parents to monitor their
environments such as in urban areas and indoors. Since          kids, personal navigation; locating a place such as nearby
the position accuracy yielded by GPS methods is better          restaurant, hotel or friends. AGPS is already used by
than other positioning technologies, most wireless carriers     many cell phone operators because of its superior
are looking at Assisted GPS (AGPS) as the solution to           positioning accuracy when compared to existing
meet the FCC criteria. A wireless network can aid a             positioning technologies such as Enhanced Observed
AGPS receiver enabling it to acquire and obtain a position      Time Difference (E-OTD) (Syrjärinne, 2001).
fix quickly under weak signal conditions (e.g. inside a
building), The purpose of this paper is to compare the          Hardware simulator provides a controlled environment to
effects of various aiding parameters such as satellite          carry out various tests. The user can conduct experiments
ephemeris or almanac, and varying timing or horizontal          in the connivance of a lab using the Radio Frequency
positioning uncertainties on AGPS signal acquisition by         (RF) simulator and can be used to carry out initial testing
using the hardware simulator and carrying out tests under       of any commercial products before rigorous field testing
different field test conditions. A SiRFLocTM evaluation         are carried out. In recent years numerous papers have
kit is used to investigate the performance of the AGPS          been published to investigate the effects of multipath and
receiver. Acquisition tests showed the importance of            or replicating the field test conditions using the simulator,
aiding data such as satellite ephemeris and good timing or      (Lachapelle et al, 2003; Thu et al, 2005). Furthermore
position accuracy under significantly weak signal               simulation tests had been carried out to investigate the
conditions, e.g. the concrete basement. Comparison of           effects of different aiding data such as different timing or
field and simulation testing showed similarities when the       position uncertainties on AGPS signal acquisition
effects of different adding data were analyzed at various       (Karunanayake at al., 2004)
simulated signal power levels or different field test
conditions thus illustrating that simulation tests which can    Field tests represent realistic situations which can be
be cheaper than field testing in terms precious resources       unpredictable because of factors such as signal blockage
such as time, cost could provide valuable first-hand            or masking and various error sources such as multipath
knowledge before conducting field tests. Limited                effects. GPS signals can undergo signal masking or
acquisition tests were carried out using the High               blockage due to natural or man-made structures, which
Sensitivity GPS receiver (unable to acquire under really        would reduce satellite availability resulting in poor
weak signal conditions, less than -140 dBm) illustrating        geometry which would degrade the user position

ION GNSS 2006, Fort Worth TX, 26 – 29 September 2006                                                                     1/12
accuracy. Materials such as glass, steel, wood or concrete
can also cause multipath effects, thus degrading the           HSGPS receivers use a combination of longer integration
positioning accuracy. Tall glass buildings in an urban or      time (coherent followed by non-coherent) as well as
sub-urban environment can block satellites from certain        massive parallel correlation to acquire and track weak
directions, therefore, reducing availability. Some GPS         signals (Rounds and Norman, 2000). Longer integration
signals may be reflected off these surfaces causing strong     times could have an adverse effect on the Time to First
specular reflections, which would require a GPS receiver       Fix (TTFF), but parallel searching of different C/A code
to acquire and or track the reflected signals.                 phases would ensure that this is not the case. HSGPS
                                                               without any assistance during a cold starts takes at least
An indoor environment presents severe challenges to GPS        30 s to download complete satellite ephemeris thus
receivers in terms of signal acquisition or tracking. GPS      resulting in longer TTFF (>30 s, open sky conditions),
signals can be attenuated by as much as 20 to 25 dB            therefore is not viable for commercial applications or
inside a building compared to outside Line-of-Sight            emergency situations, which have more stringent
(LOS) conditions (van Diggelen, 2001). Satellite               requirements on the TTFF.
availability is a major problem due to signal blockage and
the only source of GPS signals may be reflected multipath      Assisted GPS, as shown in Figure 1, gets aiding data such
signals or highly attenuated LOS signals. The inability of     as timing, approximate user position, frequency
conventional receivers to work in these types of               assistance, satellite ephemeris and almanac from a server
challenging conditions has led to the development of High      via a wireless network, for example. Assistance data is
Sensitivity GPS (HSGPS) and Assisted GPS (AGPS)                required to aid in the acquisition process, thus enabling a
receivers. AGPS is able to acquire with the help of aiding     quicker position fix resulting in a shorter TTFF.
data in these types of challenging conditions and is the       Assistance data also enhances the acquisition sensitivity
main focus of this paper                                       of the receiver and several investigations have shown that
                                                               AGPS can reach acquisition sensitivities of –150 dBm or
In recent years, field tests have been carried out using       –155 dBm (van Diggelen, 2001; Bryant et al., 2001).
AGPS receivers under many different field conditions,
e.g. Garin et al., (2002), Krasner et al., (1998), Biacs ta    AGPS methods are divided into two categories. The
al.,(2001) and Bryant et al., (2001). Results from the field   position can be computed at the mobile station (MS),
tests investigating the effects of different aiding data on    which is called MS-Based or at the network, MS-Assisted.
AGPS signal acquisition have been published in the             In the MS-Based implementation, the network would send
previous paper, Singh et al., (2005) which in fact leads us    all of the assistance data enabling the handset to acquire
to the objective of this paper where effects of aiding data    and obtain a position fix. In the MS-Assisted mode, the
using the simulator and various field testing. The             handset receives assistance such as visible satellites and
secondary purpose of this paper it to illustrate the effects   satellite Doppler data which enables the receiver to
of different aiding data on such things is acquisition         quickly acquire the satellites and to send pseudorange
sensitivity and Time to First Fix                              information to the server, which then computes the user
                                                               position (Pietilä and Williams, 2002).
                                                               AGPS is like a HSGPS receiver in hot start mode (further
A GPS receiver has to acquire signals to obtain coarse         shown in this paper), whereby it knows the approximate
estimates of satellite Doppler or code phase before it is      time, has approximate user position, ephemeris and
able to commence tracking. Longer coherent and non-            frequency thus reducing the C/A-code or Doppler bins,
coherent integration is required to acquire signals less       hence lowering the acquisition search time. Satellite
than –150 dBm (Chansarkar and Garin, 2000). The                ephemeris combined with an approximate user position
coherent integration time is limited to 20 ms due to the       can be used to predict the approximate satellite Doppler.
navigation data bits. Coherent integration can be carried      Frequency assistance will further reduce the Doppler
beyond 20 ms if the navigation data bits could be              search space. Timing and approximate position assistance
predicted, however coherent integration is further limited     can be used to predict the approximate C/A-code phase. If
due to residual frequency errors such as satellite motion,     the combined timing and position assistance is less than
receiver, clock instability and receiver induced motion.       one millisecond or 300 km then the approximate code
Integration can be carried out non-coherently to further       phase can be predicted, therefore an entire sweep of 1023
enhance the receiver sensitivity (Shewfelt et al., 2001).      code chips is not required (Kinnari, 2002). The
HSGPS is discussed next because AGPS has similar               approximate user position can also be used to determine
architecture when compared to HSGPS, however AGPS is           the visible satellites, while the almanac can be used to
able to receive aiding data from the wireless network to       predict the approximate location of the satellite
further enhance acquisition sensitivity and lower the          constellation or satellite visibility.

ION GNSS 2006, Fort Worth TX, 26 – 29 September 2006                                                                  2/12
                                                                    4) Effects of different timing and position
                                                                       uncertainty on AGPS (Karunanayake, 2005)

                                                                The acquisition tests using AGPS receiver with different
                                                                starts such as hot or warm start were carried out by
                                                                keeping all the satellite power at nominal (-130 dBm) for
                                                                fifteen minutes so that complete satellite almanac and the
                                                                satellite ephemeris can be downloaded. While carrying
                                                                out acquisition tests in hot, warm and cold start, the
                                                                AGPS receiver did not received any aiding data from the
                                                                reference receiver. Acquisition results were quantified by
                                                                using factors such as TTFF (normalized with respect to
               Figure 1: The AGPS System                        the maximum TTFF), positional accuracy, number of
                                                                satellites used in the position solution and simulator
                                                                power. The single-point position results were obtained by
SIMULATION TESTS                                                processing the raw pseudorange data from the receiver
                                                                using the C3NAVG2 software (Petovello et al., 2000).
Acquisition tests as illustrated in Figures 2 and 3 were        The C3NAVG2 software, developed by the PLAN group
carried out by using the Spirant GSS 6560 Simulator. The        at the University of Calgary uses the least squares
simulator consists of a PC (software), which is connected       algorithm to estimate the epoch-to-epoch position, which
to two hardware RF simulators. The hardware simulator           is very suitable for our analysis, since no filtering is
will be referred to as vehicles, where each vehicle is          performed. The results are shown in Figures 4, 5, 6, 7, 8
capable of simulating a maximum twelve satellites where         and 9. Table one shows the position results of the AGPS
parameters such as satellite power can be changed.              receiver at different power levels, while Table 2 shows
Acquisition tests were carried out by changing the              the acquisition sensitivities for different tests.
simulator power until the receiver stops acquiring. The
receivers used for the tests were SiRFLocTM AGPS and
SiRFXTracTM HSGPS receivers ( The
SiRFLocTM AGPS is a multimode receiver capable of
receiving aiding data from various wireless networks such
as CDMA or GSM. The SiRF receivers are based on the
StarII architecture and will be referred to AGPS and
HSGPS from herein. The test receivers were connected
vehicle one, while the reference receiver is connected to
vehicle two which had nominal signals (-130 dBm) and
provided reference data such as timing with an
uncertainty of 125 µs, approximate user position with
horizontal and vertical uncertainties of 5 km and 150 m
respectively, satellite ephemeris, almanac and frequency                  Figure 2: Simulator Setup Diagram
assistance to the AGPS receiver. The simulator power is
referenced to –130 dBm and can go down to –150 dBm.
The simulator power was decreased further by using a 7
dB attenuator. Acquisition tests were started at –130 dBm
and the power was decreased by one or two dB until the
receiver stopped acquiring. The test receivers tried to
acquire and obtain thirty position fixes at different power
levels. The measurement errors such as multipath
Ionospheric or Tropospheric were not simulated during all
the tests, therefore showed effects of thermal noise that is,
decrease in signal power results in higher thermal noise.
Simulation test objectives are discussed next.

Simulation Test objectives:

    1) Acquisition sensitivity of AGPS and HSGPS                               Figure 3: Simulator Setup
    2) Effects hot warm and cold starts on AGPS
    3) Ephemeris or Almanac on AGPS

ION GNSS 2006, Fort Worth TX, 26 – 29 September 2006                                                                  3/12
                                                        Figure 7: Comparison of Hot, Warm and Cold start
   Figure 4: Comparison of AGPS test with Default                     for AGPS Receiver
       Aiding,, Hot start and without Almanac

                                                       Figure 8: AGPS with different Timing uncertainties at
Figure 5: Comparison of AGPS Test with Warm Start      Various Power Levels (Source: Karunanayake, 2005)
              and without Ephemeris

                                                       Figure 9: AGPS with different Position uncertainties
                                                                  (Source: Karunanayake, 2005)
 Figure 6: Comparison of AGPS with Cold start and

ION GNSS 2006, Fort Worth TX, 26 – 29 September 2006                                                     4/12
 Table 1: AGPS Position Results with all Aiding Data              showed that warm start or AGPS without ephemeras had
            using Least Squares Solution                          the similar results suggesting that AGPS (without
                                                                  assistance) in warm start is nothing more than a AGPS
                   Five Position              Thirty Position     without ephemeris assistance. The test also illustrated the
                                                                  ephemeris can be used to enhance acquisition sensitivity
                          Fixes                     Fixes         that is an improvement 11 dB was achieved with
   Simulator                                                      ephemeris aiding which in fact suggests that it is difficult
                    2D       Mean            2D                   to demodulate the navigation data bits under really weak
     Power                                            Mean # of   signal conditions (< -142 dBm). The simulation test using
                  RMS         # of         RMS                    AGPS and HSGPS in cold start had similar results, when
     (dBm)                                                SVs     aiding is not available (wireless network outage) AGPS
                  Error       SVs          Error                  receiver performs like a high sensitivity receiver therefore
                                                                  performance is not severely degraded under adverse
                   (m)                      (m)                   signaling conditions. The AGPS comparing hot (same as
                                                                  AGPS in assisted-mode), warm and cold start, showed
       -130         3.6           8.2        3.8            7.9   that hot start performed the best in terms of acquisition
                                                                  sensitivity and TTFF since the receiver the satellite
       -136         6.1           8.0        3.4            8.2   ephemeris, acquisition search space (C/A code and
                                                                  Doppler) is reduced with approximate GPS time and
       -142        10.9           8.5       14.5            7.8   location is also known fairly accurately thus reducing the
                                                                  acquisition search time. A receiver without any satellite
       -148        41.1           4.5       45.5            4.7   ephemeris would take up to 30 s resulting in longer TTFF
                                                                  which could be observed when the warm start was
       -152        37.0           4.0       34.8            4.0   compared with hot start. Finally the position results
                                                                  showed that horizontal position accuracy degraded with
                                                                  lower signal power, higher thermal noise would result in
     Table 2: Acquisition Sensitivities for Different             larger measurement errors, the numbers of satellites used
                   Simulation Tests                               decreased with lower signal power, thus resulting in poor
                                                                  geometry which degraded the horizontal position
                                        Acquisition Sensitivity   accuracy. The acquisition tests that were carried out using
         Receivers                                                varying time or position uncertainties by Karunanayake,
                                                                  (2005) have shown increasing the timing or horizontal
             Cold Start                          -140
                                                                  position uncertainty at -140 dBm increased the TTFF,
                                                   -142           while decreasing the simulator power with constant
                                                                  timing uncertainty also led to longer TTFF, longer
             Hot Start                             -152
                                                                  integration time is required to acquire weaker GPS
   SiRF       Assisted
                                                   -153           signals.
   AGPS        mode
                No                                                FIELD TESTS
                No                                                The AGPS receiver was used to carry out majority of the
             Ephemeris                                            acquisition tests under different field test conditions. The
      SiRF HSGPS                                   -140           test methodology would remain the same regardless of the
                                                                  environment that was used to carry out the tests. The test-
                                                                  setup is shown in Figure 8 shows the AGPS receiver and
The simulations tests have shown that AGPS had better             the reference receiver, where the reference receiver is
sensitivity (13 dB) when compared to the HSGPS                    connected via a 30 m cable to a reference antenna which
receiver. The higher sensitivity of the AGPS receiver is          has clear LOS signals. The reference receiver provides
due to the aiding data which as have been demonstrated            aiding data such as satellite ephemeris, almanac,
by carrying out simulation tests using different types of         approximate GPS time, and approximate user position
starts and or aiding data (no ephemeris or almanac). The          with horizontal and vertical uncertainties. The AGPS
tests with AGPS using all the aiding data, without                receiver is connected to the micopatch antenna. Each test
almanac or hot start had similar performance in terms of          is carried out from a cold start, GPS time, satellite
TTFF and acquisition sensitivity suggesting that almanac          ephemeris or almanac. The commands are given by using
is not required since it provides a coarse estimate of            the data-logger software. During each test, more than
satellite orbital parameters, while the test with hot start       twenty trials were carried out and 30 position fixes were
showed that AGPS is essentially a HSGPS receiver that             obtained for each trial. After 30 position fixes, the AGPS
restarted in the hot start mode. Simulation tests also

ION GNSS 2006, Fort Worth TX, 26 – 29 September 2006                                                                      5/12
receiver would restart from a cold start to obtain the         AGPS signal acquisition. The field tests were carried out
position fix for the next trial. The acquisition test is       in the sub-urban environment, residential garage, speed-
carried out in the similar manner using the HSGPS              skating track and the concrete basement. The environment
receiver however it did not receiver any aiding data.          reflects the realistic lifelike scenarios where the mobile
Field test objectives are discussed next.                      devices can either used for emergency E-911 calls,
                                                               personal navigation or other commercial LBS
Field Test Objectives:                                         applications. The results were analyzed by using factors
                                                               such as TTFF, position accuracy (obtained using
         1) Investigate the acquisition performance of         C3NAVG2) and C/N0, which was obtained from the
            the HSGPS receiver                                 receiver.
         2) Determine the effects of satellite ephemeris
            or almanac on AGPS signal acquisition              The sub-urban test was carried out at surveyed point at the
         3) Investigate the effects of different timing or     University of Calgary campus on October 9, 2004. The
            different horizontal position uncertainties on     test site is shown in Figures 11 and 12. There was a tall
            AGPS signal acquisition.                           glass building on the east side, a smaller concrete building
                                                               on the west and a glass walkway to the north, which could
During the acquisition tests, the approximate user position    cause some satellite blockages. The southern side is
of the receiver was set to one of the surveyed points at the   unhindered with some coniferous trees in the south west
University of Calgary campus. The horizontal position          side.
uncertainty was kept fixed to 5 km when the time
uncertainty was changed, while the timing uncertainty
was kept the same 125 µs while the horizontal position
uncertainty was changed. The AGPS receiver would
continue to receive almanac and ephemeris for varying
time or horizontal position uncertainty tests. The timing
and horizontal position uncertainty was set to 125 µs and
5 km for tests that were carried out without ephemeris or
almanac. The vertical position uncertainty was kept the
same at 150 m for all the tests.

                                                                          Figure 11: Sub-Urban Test Setup

               Figure 10: Field Test Setup


The tests were conducted under many different conditions
which would presented many challenges such as signal
blockage which could reduce satellite availability,
multipath effects, high level of signal attenuation which
results in weak signals (< -150 dBm) or low C/N0.
Acquisitions tests were carried out in the suburban
environment, residential garage, inside a speed-skating
track and a concrete basement. The different test sites
were chosen to determine the effects of various factors on
                                                                         Figure 12: Sub-Urban Surrounding

ION GNSS 2006, Fort Worth TX, 26 – 29 September 2006                                                                   6/12
                                                              Calgary campus on December 13, 2004. The Indoor
The acquisition tests were carried out in a residential       Speed skating track is made of made of concrete beams
garage (located within 5 km from the university) on           with dimensions 198 by 80 m and roof is constructed of
December 9, 2004. The garage setup as shown in Figures        corrugate steel (interior) and porcelain panels (exterior).
13 and 14 was underneath the living room of a house and       Field tests were carried out near the window so that the
the walls are made up of wood and concrete. The door          reference receiver could be connected to the antenna
facing the East side was made up of wood, while the wall      which was placed near the window. The AGPS receive
facing the South side was partially constructed of wood       was connected to the microstrip patch antenna which was
and the remaining two walls facing West and North were        placed at a surveyed point inside the building. This
made of concrete. The garage door was closed during all       environment would highly attenuate signals from curtain
acquisition tests and the reference antenna, which was        directions due to concrete walls, while strong signals
connected to the reference receiver, was located outside      could enter from the window and there could be reflected
the garage. The AGPS receiver was connected to a              signals due to structures such as the corrugated roof.
microstrip patch antenna, which was placed at a surveyed
point inside the garage.

                                                                        Figure 15: Speed-skating Test Setup
       Figure 13: Residential Garage Test Setup

                                                                    Figure 16: Speed-skating Test Surrounding

      Figure 14: Residential Garage Surrounding               The acquisition test as shown in Figures 17 and 18 was
                                                              carried out in a concrete basement which is located at the
The acquisition tests were carried out in a speed-skating     University of Calgary campus on November 30, 2004 The
track (Figures 15 and 16) which is located at University of   basement represented an extremely difficult environment
                                                              with a lot of satellite blockages and highly attenuated

ION GNSS 2006, Fort Worth TX, 26 – 29 September 2006                                                                 7/12
signals. There is a door located on the North side, which     obtain a position fix in the speed-skating track and the
was made of wood and two small windows were located           concrete basement or the speed-skating track because of
on the North West side. The remaining three side walls        really weak signals making it impossible to download the
were made of reinforced steel concrete. Similar to the        satellite ephemeris. The TTFF was for the HSGPS
previous tests, the AGPS receiver was connected to a          receiver was ten times more than the AGPS receiver on
microstrip patch antenna which was placed at a surveyed       the roof and in the residential garage. The position results
point inside the concrete basement, while the reference       are shown in Tab\le 3. The limitations using HSGPS
antenna was kept outside.                                     which have been shown by the field and simulation tests
                                                              have led to the development of AGPS and have been used
                                                              to carry out numerous tests with different aiding scenarios
                                                              and are discussed in the next.

                                                                Table 3: Position Results for the HSGPS Receiver
                                                                          using Least Squares Solution
                                                                                  Five Position    Thirty Position
                                                                                      Fixes             Fixes
                                                                                 2D      Mean #     2D     Mean #
                                                                 Type of        RMS                RMS
                                                                               Error        of     Error      of
                                                                                 (m)                (m)
                                                                                         Satellite         Satellite

                                                                                   5.9       8.5        5.5       8.7
       Figure 17: Concrete Basement Test Setup                                    19.9       5.5       14.3       4.3

                                                              FIELD TESTS - ASSISTED GPS

                                                              The acquisition tests showed that the AGPS receiver was
                                                              able to perform signal acquisition without ephemeris in
                                                              the sub-urban and residential garage. The TTFF was ten
                                                              times more than with time and position aiding of 125 µs
                                                              and 5 km in sub-urban environment while it was six times
                                                              more for the garage test when compared with the same
                                                              scenario. The acquisition tests could not be carried out in
                                                              the other two environments such as the speed-skating
                                                              track and the concrete basement without ephemeris, it was
                                                              difficult to demodulate the navigation data bits in the two
                                                              environment hence making it impossible to acquire
   Figure 18: Concrete Basement Test Surrounding              without ephemeris, the results also illustrated the
                                                              limitations of the HSGPS receiver which was unable to
                                                              acquire or obtain a position fix in the two environments
FIELD TESTS - HIGH SENSIVITY GPS (HSGPS)                      because HSGPS (cold start) is required download the
                                                              satellite ephemeris before it can carry out signal
Limited field tests were carried out using the HSGPS          acquisition. The acquisition tests with almanac assistance
receiver in the residential garage and using was one of the   had the same TTFF when compared with the time and
antennas on the Calgary Centre for innovative                 position assistance of 125 µs and 5 km for all the four
Technologies (CCIT) roof. The results obtained from the       field tests. If the receiver without ephemeris is required to
roof would similar to the sub-urban environment be            download it which can take up to thirty seconds thus
similar in terms of the TTTF because it had nominal GPS       prolonging before performing signal acquisition resulting
signals. The HSGPS receiver could not acquire and or          in longer TTFF. Satellite almanac unlike satellite

ION GNSS 2006, Fort Worth TX, 26 – 29 September 2006                                                                    8/12
ephemeris provides coarse information for such things as
satellite orbits therefore is not required for signal
acquisition. The position results for ephemeris or almanac
assistance is given in Tables 4 and 5. The tests showing
results with different timing or position uncertainties are
shown in Figures 19 and 20 while Tables 5 and six shows
the position results and C/N0 are shown in Tables 5 and 6.

Table 4: Position Results for Ephemeris and Almanac
Aiding in different Field Test conditions Aiding Using
                 Least Squares Solution
                               Five           Thirty
                             Position        Position
                               Fixes          Fixes
  Field Tests Sites and
                             2D      Me     2D      Mea
    Aiding Scenarios
                           RMS       an   RMS       n#
                           Error # of Error          of
                            (m)      SV    (m)      SV
Sub-Urban       Almanac     41.4     6.7   35.9     7.2         Figure 20: TTFF and Number of Satellites using
               Ephemeris                                           different Horizontal Position Uncertainties
                            49.4     7.1   40.4     6.7
Residential     Almanac     25.9     5.3   25.2     6.2       Table 5: Position Results using Least Squires Solution
  Garage       Ephemeris    21.3     5.2   18.5     6.4       using Ephemeris Almanac, Time (125 µs) and Position
  Speed-        Almanac     57.9     3.4   55.4     3.3        Aiding (5 km) under Different Field Test Conditions
   Track       Ephemeris     X        X     X        X                          Five Position       Thirty Position
 Concrete       Almanac     63.2     4.1   56.1     3.4
 Basement Ephemeris          X        X     X        X                                Fixes                Fixes

                                                               Field Test       2D                                 Mean
                                                                                          Mean #    2D RMS
                                                                  Sites        RMS                                 # of
                                                                                          of SVs    Error
                                                                              Error                                SVs

                                                                 Roof           5.6           8.2    5.8           7.9

                                                                               60.8           5.6    60.1          5.7

                                                                               18.8           5.1    23.5          5.0


                                                                               45.7           3.7    37.5          3.8

   Figure 19: TTFF and Number of Satellites using
                                                                               61.7           5.3    68.4          3.2
        different Precise Time Uncertainties                   basement

ION GNSS 2006, Fort Worth TX, 26 – 29 September 2006                                                                9/12
The trends between the timing and position assistance          greater uncertainty in the C/A code phase, while the
showed similar results with the number of satellites           horizontal position uncertainty will result in greater
tracked and the TTFF (larger timing or position                uncertainty in C/A code phase and Doppler (1 Hz/ km,
uncertainty results in longer TTFF). At the risk of            van Diggelen, 2001)
repeating ourselves the results from the test with varying
time uncertainty will be discussed next, when two
different environment are compared using time aiding of        COMPARISON OF SIMUALTION & FIELD TESTS
125 µs. The TTFF of the roof test was showed similar
trend when compared to the sub-urban environment, both         Simulation tests were carried out in a controlled
tests sites had nominal signal conditions (39 to 45 dB-        environment where every channel had same power unlike
Hz), position accuracy was worse in the sub-urban              field test conditions where the signal power varied due to
environment which could be due to such things as; lower        various factors such signal attenuation signal blockage or
number of satellites resulting in poor geometry and            reflected signals from different surfaces, no measurement
multipath effects (glass building would cause strong           errors were simulated thus results could not be compared
specular reflections). The residential garage had              using position results, however results using two
attenuated signals (30 dB-Hz) had TTFF which was five          methodologies showed similar trends in TTFF and
times longer than the sub-urban environment; however           acquisition sensitivity. The results from the simulation
the AGPS receiver had better position accuracy due to          and field tests had similar results when tested with
factors such as better geometry and multipath from             different aiding scenarios for e.g. tests with almanac
diffuse sources (sources could be wood or concrete) are        aiding showed similar results when compared to various
not severe as those from specular reflective sources (see      field tests, HSGPS receiver had similar results in terms of
Section Error! Reference source not found.). Similarly         TTFF and sensitivity, that is the receiver was unable to
the speed-skating environment with signal strength of 25       acquire beyond -140 dB which is further reflected with
dB-Hz had TTFF which was twice as long when                    the receivers inability to acquire under really weak field
compared the residential garage and had poorer position        test conditions such as concrete-basement or speed-
accuracy which could be due to factors such as poor            skating track. The tests without ephemeris assistance
geometry or multipath, specular reflections that could due     could not be carried beyond -142 dBm, which were
to sources such as corrugated roof. Although the concrete      further confirmed by field tests where the AGPS was
basement and the speed-skating track had similar signal        unable to acquire without ephemeris in weak signal
conditions that is highly attenuated signals, the TTFF was     conditions; in fact this scenario is similar to the HSGPS in
twice as long for the concrete basement because the test in    a cold or warm start where it has to download satellite
the speed-skating was carried out near the window where        before performing signal acquisition. The tests with
there was one strong satellites (PRN 5) resulting in lower     varying timing and varying position accuracies had
TTFF. Simulation tests in Karunanayake, (2005) have            similar results (longer TTFF) using two methodologies
shown the importance of initially acquiring the strong         where greater uncertainty in time or horizontal position
satellite which can then be used to aid in acquiring the       under weak signals (-140 dBm or weak field conditions).
remaining weaker satellites. Once the first strong satellite   Field tests or simulation tests using different simulator
is acquired it can be used to provide such things as           power levels showed similar trends when compared to
accurate GPS timing, clock bias which can be used to           tests that were carried out under various signal conditions
acquire the remaining satellites. The results have shown       that is tests under weaker signal conditions required
that it takes a longer time acquires signal in weaker signal   longer integration time thus resulting in longer TTFF.
conditions; weaker signals implies longer search time or       Simulator tests were carried out under the simplest
integration time thus longer TTFF, while increasing the        conditions without any error sources to get the feel for
timing or position uncertainty will result in larger search    technology or categories them in terms of factors such as
range in the C/A code Doppler space, with greater              acquisition sensitivities and TTFF, further tests can then
uncertainty in weak signals will result in really long TTFF    be carried under various field conditions (where people
which can be observed from the results that were obtained      may use them) to investigate their performance which was
from the tests that were carried out in the concrete           in fact one of the goals of this paper.
basement test and the speed-skating track. Wireless
assistance data such approximate C/A code-phase or C/N0
can be used to narrow the acquisition search space, hence      CONCLUSIONS & RECOMMENDATION
quicken the acquisition search process, once the C/N0 is
known, certain predefined value of integration time can        Simulation tests illustrated the importance of aiding data
be used to carry out the correlation process which would       such ephemeris in lowering TTFF and acquisition
shorten the acquisition search process. Typically the value    sensitivity (11 dB improvement with ephemeris aiding).
of integration time can be determined after acquiring the      The simulation and field tests also showed that satellite
strongest satellite. Longer timing uncertainty will lead       almanac is not required for acquisition because it provides

ION GNSS 2006, Fort Worth TX, 26 – 29 September 2006                                                                  10/12
coarse estimate of satellite orbital parameters so is not       Diggelen, F (2001), Global Locate Indoor GPS Chipset &
required and should not have any effect on TTFF, this           Services, Proceedings ION GPS-2001, September 11-14,
was further confirmed by various field tests. The tests         Salt Lake City UT, pp. 1515 1521.
using varying timing and position uncertainty showed the
importance of good timing or position accuracy especially       Eisfeller, B. (2004), D. Gänsch, S. Müller, and A. Teuber,
under weak simulated or field signal conditions.                Indoor Positioning using Wireless LAN Radio Signals,
                                                                Proceedings of ION GPS-2004 September 21-24, Long
                                                                Beach, CA, pp. 1936-1947.
All the field tests were carried out under static conditions
thus in future further field tests needs to done under          Federal Communications Commission – Enhanced 911
kinematic conditions (Karunanayake et. al, 2005).               (FCC-E911) Mandate (2000)
Simulation tests were carried out without any errors, 
further tests is required under simulated conditions that       blic_Notices/1999/da992130.html, (Accessed on August
replicated field tests sites, Thu (2006). The tests also        8, 2005)
showed that AGPS had very long TTFFs under really
weak signal conditions such as the concrete basement,           Garin, L. J, M. S. Phatak, and H. Gehue (2002), Indoor
requiring further investigation on the possibility of hybrid    GPS Test Methodology and Indoor Performance Tests,
technologies such as AGPS with Wireless Local Area              Proceedings of ION GPS-2002, September 24-27,
Network (WLAN) (Eisfeller et al., 2004) which is similar        Portland, OR, pp. 192-199.
in concept to hybrid technologies such as Cellular
(CDMA network)/AGPS that is used by Qualcomm Inc to             Hu, T., G. Lachapelle and R. Klukas (2005), Indoor GPS
increase position solution availability (Biacs et al., 2001).   Signal Replication Using aHardware Simulator,
                                                                Proceedings of ION GNSS 2005, Long Beach, CA,
                                                                September 13- 16, in press.
                                                                Karunanayake, M. D M.E Cannon, G. Lachapelle and G.
I would also like to thank my co-supervisors Dr Elizabeth       Cox (2004),, Evaluation of Assisted GPS (AGPS) in
Cannon Dr Richard Klukas for their continued guidance           Weak Signal Environments Using a Hardware Simulator,
and financial support. We would like to thank SiRF              Proceedings of ION GPS-2004 September 21-24, Long
Technology Inc. and their staff for providing the AGPS          Beach, CA, pp. 2416-2426.
and reference receivers and their technical support, which
was very valuable for this research. We would also like to      Karunanayake, M.D., M.E. Cannon, G. Lachapelle and G.
thank Merlin Keillor, Judy Smith and staff at the               Cox (2005), Effect of Kinematics and Interference on
Disability Resource Centre for their continued support.         Assisted GPS (AGPS), Proceedings of ION NTM 2005,
Dharshaka Karunanayake (my dear friend and AGPS                 San Diego, CA, January 24-26, pp. 1071-1081.
colleague) and fellow colleagues in the PLAN group are
thanked for their encouragement and support throughout          Karunanayake, M. D (2005), Hardware Simulator
this research.                                                  characteristics of Assisted GPS, M.Sc. Thesis, The
                                                                University of Calgary, UCGE Report 20231
                                                                Kinnari, T. (2001), Accurate Time Transfer in Assisted
Bryant, R, S. Dougan and E. Glennon (2001), GPS                 GPS, MSc Thesis, Tampere University of Technology,
Receiver Algorithms and Systems for Weak Signal                 Finland.
Operation, Proceedings of ION GPS-2001, September 11-
14, Salt Lake, UT, pp 1500-1510                                 MacGougan,     G. (2003), High-Sensitivity       GPS
                                                                Performance Analysis in Degraded Signal Environments,
Biacs, B., G. Marshall, M. Moeglein and W. Riley (2002),        M.Sc.    Thesis,   The     University   of    Calgary
The Qualcomm/SnapTrack Wireless-Assisted GPS          
Hybrid Positioning System and Results from Initial
Commercial Deployments, Proceedings of ION GPS                  Petovello, M., M.E. Cannon and G. Lachapelle (2000),
2002, Portland, OR, September 24-27, pp. 378-384.               C3NAVG2 Operating Manual, Department of Geomatics
                                                                Engineering, University of Calgary, May.
Chansarkar, M and L. Garin (2000), Acquisition of GPS
Signals at Very Low Signal TO Noise Ratio, Proceedings          Pietilä, S and M. Williams (2002), Mobile Location
of ION NTM 2000, Anaheim CA January 26-28, pp. 731-             Application and Enabling Technologies, Proceedings of
737.                                                            ION GPS-2002 September 24-27, Portland, OR, pp,

ION GNSS 2006, Fort Worth TX, 26 – 29 September 2006                                                                 11/12
Rounds, S. F and C. Norman (2000), Combined Parallel
and Sequential Detection for Improved GPS Detection,
Proceedings of ION NTM 2000, San Diego CA January
26-28, pp. 368-372.

Shewfelt, J.L, R. Nishikawa, C. Norman and G.F. Cox
(2001). Enhanced Sensitivity for Acquisition in weak
Signal Environments through the use of Extended Dwell
Times, Proceedings ION GPS-2001, September 11-14,
Salt Lake City, UT, pp.155-162.

Singh, S., M. E. Cannon, R. Klukas and G. Cox (2005),
Field Test Assessment of Assisted GPS and High
Sensitivity GPS Receivers under Weak/Degraded Signal
Conditions, Proceedings of ION GNSS 2005, Long
Beach, CA, September 13-16, pp .

Sudhir, N S, C. Vimala and J.K. Ray (2001), Receiver
Sensitivity Analysis and Results, Proceedings ION GPS-
2001, September 11-14, Salt Lake City, UT, pp. 1420-

Syrjärinne, J. (2001), Studies of Modern Techniques for
Personal Positioning PhD Thesis, Tempere University of
Technology, Finland.

Syrjärinne, J. and T. Kinnari (2002), Analysis of GPS
Time-Transfer Accuracy in GSM and UMTS Networks
and Possibilities to Improve Sensitivity, Proceedings of
ION GPS-2002, September 24-27, Portland, OR, pp. 184-

ION GNSS 2006, Fort Worth TX, 26 – 29 September 2006       12/12