Satellite Positioning (GE 703)

BGE IV/I

Chapter 4: System Biases and Error [8 hrs]

▪ Multipath
▪ Timing and Orbital Biases
▪ Troposphere
▪ Ionosphere

Asst. Prof., Er. Bikash Sherchan 1

 Sources of Errors
❑ Various sources of errors – affect GNSS
performance
❑ Sum of all systematic error – referred to as range
bias
❑ Observed range – without removal of biases –
called biased range
❑ Principal sources:
❖ Ephemeris Error (Orbital Error)
❖ Clock Biases
❖ Atmospheric Biases
❖ Receiver Noise
❖ Multipath Effect
2
 Sources of Errors
❑ Other Sources:
❖ Selective Availability (S/A)
❖ Anti-spoofing (A/S)
❖ Random Observation Errors (enexplainable and
unpredictable time variation)
❑ These errors – associated with both absolute and
differential positioning
❑ Many of these errors – either eliminated or
significantly minimized – observations done in
differential mode
❑ Because – the same errors being common to
both receivers during simultaneous observations3
 Sources of Errors
❑ Both code and phase pseuduranges – affected by
the systematic and random noises
❑ Error sources – can be classified into 3 groups:
❖ Satellite related errors
❖ Propagation related errors and
❖ Receiver related errors
❑ Some of the systematic errors – possible to
model – can be eliminated by appropriate
combinations of observations
❑ Double differenced pseudoranges – free of
systematic errors related to satellite and
receivers 4
 Sources of Errors
❑ Errors caused by ionospheric refraction –
virtually eliminated by combination of dual
frequency data
❑ Measurement noise, satellite biases and
propagation error – combined in the UERE
(User Equivalent Range Error)
❑ Simply called as URE (User Range Error)
❑ UERE – transmitted via navigation message
❑ UERE in combination with DoP factor – user
can estimate the accuracy
5
 Sources of Errors
❑ Range Biases

Source Effect
Satellite Clock Bias
Orbital Errors
Signal Propagation Ionospheric Refraction
Tropospheric Refraction
Receiver Antenna Phase Center Variation
Clock Bias
Multipath
6
 Ephemeris or Orbital Errors
❑ Errors in prediction of satellite position
❑ Typically size of satellite ephemeris error < 8 m
(95% confidence)
❑ Satellite dependent – very difficult to completely
correct and compensate
❑ Many forces acting on satellite while it is
orbiting is difficult to measure
❑ More accurate data – available later stage for
post processing – not practical for real-time
positioning
7
 Ephemeris or Orbital Errors
❑ Satellites – travel in a very precise and well-
known orbits
❑ Like dampening of satellite clocks – orbits too
vary a small amount – results in significant error
in position
❑ Ground control system (GCS) – monitors
satellite orbit continuously
❑ As satellite orbit changes – GCS sends
corrections to the satellite – ephemeris is
updated
8
 Ephemeris or Orbital Errors
❑ Even with the correction from GCS – still small
errors remains in the orbit – results up to ± 2.5 m
positional error
❑ One way of compensation – download precise
ephemeris information from SBAS or PPP service
provider
❑ Another way of rectification – differential GNSS or
RTK receiver configuration
❑ Errors present in the orbit is in the ratio
𝛿𝑏 𝛿𝑆
=
𝑏 𝑅𝑠
9
 Ephemeris or Orbital Errors
❑ Errors present in the orbit is in the ratio
𝛿𝑏 𝛿𝑆
=
𝑏 𝑅𝑠
where 𝛿𝑏 = error in baseline b
𝛿𝑆 = error in satellite orbit
𝑅𝑠 = satellite range
❑ Specification – orbital error shall not exceed 3.7 m
❑ Not always possible
❑ Using broadcast ephemeris – orbital error ranges
between 10 to 20 m
❑ Substantially eliminated using differential techniques
10
 Clock Biases
❑ GNSS performance – relies heavily on accurate time
measurement
❑ Satellite clocks – rubidium and cesium
❑ Rubidium - accurate to 1 part in 1012 – lose 1 sec
only every 30 thousand years
❑ Cesium - accurate to 1 part in 1013 – lose 1 sec only
every 300 thousand years
❑ Hydrogen Maser - lose only 1 sec every 30,000
thousand years
❑ Most of the receiver clocks – quartz
❑ Quartz - accurate to 1 part in 108
❑ Time offsets between the satellite and receiver clocks
❑ As a result – erroneous range
11
 Clock Biases
❑ Range error - as a result of time offset – approximately
𝑅𝐸 = 𝜏𝑜 𝑐
where 𝑅𝐸 = Range Error due to Clock
Instability
𝜏𝑜 = Time Offset
c = Speed of Light (299,792,458 m/s)
❑ Atomic clocks in the satellites – very accurate –
however they drift by a small amount
❑ Unfortunately this small inaccuracy – results in
significant error in position
❑ For time offset of 1 microsecond (10-6 s) ≈ 300 m range
error
12
 Clock Biases
❑ Excessive temperature variation – variation of the
satellite clock from GNSS time
❑ Careful monitoring – assess the amount of drift –
included in the broadcast message
❑ Error due to this drift - substantially eliminated if
the user uses the same data
❑ Satellite clocks – monitored by GCS – using even
more accurate clocks
❑ For more accurate position – receiver need to
compensate for clock error
13
 Clock Biases
❑ Timing error – exist at the receiver end as well
❑ Reason – receiver clock in being compatible, not in
the same time system as the satellite clock
❑ Range measurement – contaminated
❑ Such error – can be evaluated using 4 satellites
or can be cancelled using differencing techniques
❑ GNSS receivers – equipped with quartz crystal
clock – relatively cheaper and compact
❑ Power requirement – very low – long life-span

14
 Clock Biases
❑ Reliability: 1 part in 108 to a maximum of about 1 part in
1010 – drift ≈ 0.1 ns in 1 s
❑ However, quartz clocks – not stable as the atomic
standards
❑ They are more sensitive to temperature changes, shock
and vibration
❑ Receiver clocks – relatively stable – but not
sophisticated at in the satellite clocks – not desired as
well
❑ If they are made to great stable – cost will be extremely
high
❑ Satellite clock stability – not necessary in receiver
clock – 4 unknowns (x, y, z, t) – solved with 4
observations
15
 Atmospheric Refraction
❑ Atmosphere – sub-divided into regions
❑ Sub-regions with similar composition and properties –
called spheres
❑ Boundary layer between spheres – called pause
❑ 2 spheres of concern in GNSS – troposphere and ionosphere
❑ Troposphere – lowest part of the atmosphere – exist up to
10 to 20 km in altitude
❑ Tropopause – separates the troposphere from the
stratosphere
❑ Stratosphere – goes up to 50 km
❑ Combined refraction (stratosphere, tropopause and
troposphere) - tropospheric refraction
16
 Atmospheric Refraction
❑ Several other layers – above 50 km
❑ Most concerned in satellite survey – ionosphere
❑ Ionosphere – 50 km to 1000 km above the earth
❑ Satellite signals – refracted through ionosphere and
troposphere – range error
❑ Error due to atmospheric refraction – dealt in two
parts: ionospheric and tropospheric delays

17
Atmospheric Refraction

© NOAA, 2014 18
 Atmospheric Refraction – Ionospheric Delay
❑ As the name suggests – composed of ions – positively
charged atoms and molecules and free electrons
(negatively charged)
❑ Free electrons – delay the signal propagation
❑ Number of ions – depends on the Sun’s UV radiation
❑ Solar flare activity – aka space weather – increase the
number of ions dramatically
❑ Therefore – reason for concern – avoid survey works
during such activity
❑ Solar flare follows a peak variation of 11 years – max
in 2011, min in 2005
19
 Atmospheric Refraction – Ionospheric Delay
❑ Ionospheric refraction – largest error in satellite
positioning
❑ Important to assess – space weather during GNSS
survey
❑ Also depends on – satellite geometry and the time of
the day
❑ Can vary – 40 to 60 m during the day – 6 to 12 m
during the night
❑ Frequency dependent – dual frequency receivers –
capable to resolve errors due to ionospheric refraction
❑ Both signals – measured continuously and differenced
– delay reduced to less than 5 m 20
 Atmospheric Refraction – Ionospheric Delay
❑ Single frequency receivers – whether in absolute or
differential mode – rely on ionospheric models to
reduce the effect of ionospheric refraction
❑ Due to varying nature of ionospheric delay – models
are not effective compared to using multi-frequency
receivers
❑ Ionospheric conditions – very similar within a locality
– base and rover experience similar delay
❑ Therefore – differential or RTK mode – effective to
compensate ionospheric delay
❑ Satellite near zenith – least amount of ionospheric
delay 21
 Atmospheric Refraction – Tropospheric Delay
❑ Lowest layer of the atmosphere
❑ More variable than ionosphere – not frequency
dependent
❑ Being close to the ground – atmospheric parameters
(temperature, pressure, humidity, etc.) – easily
measured –integrated effect can be modeled – along
the line of sight
❑ If conditions are similar – base and rover experience
similar tropospheric delay
❑ Allows differential and RTK mode over short baseline
– completely eliminate tropospheric delay
22
Atmospheric Refraction – Tropospheric Delay

Tropospheric effect (© Sickle, 2015) 23

 Atmospheric Refraction – Tropospheric Delay
❑ Over larger baselines, measurements can be taken –
used in appropriate model – reduce the error by 95%
❑ Tropospheric condition causing refraction of GNSS
signal – can be modeled by measuring dry and wet
components
❑ Signals in the L-band – not dispersed by the
troposphere – but refracted due to moisture in the lower
atmosphere
❑ Dry component – best approximated by
𝐷𝐶 = 2.27 × 0.001𝑃𝑜 [𝑚]
where 𝐷𝐶 = Dry term range contribution in Zenith
Direction
𝑃𝑜 = Surface Pressure [mb] 24
 Atmospheric Refraction – Tropospheric Delay
❑ Wet component – more difficult to approximate
❑ Apart from surface conditions – depends on various
factors such as water vapor content, altitude, and angle
of signal path
❑ No well-correlated model – for approximation of wet
component

25
 Multipath
❑ Error due to signals arriving from more than one
path – multipath
❑ Results in inaccurate positions
❑ Cause of multipath – reflecting surfaces like
metal, buildings or ground – time delay –
increased range
❑ Different geometric situations – no general
model to deal with multipath effect
❑ Its influence estimated – using combined L1 and
L2 code and carrier phase measurements
26
 Multipath

Direct signal

© Ghilani and Wolf, 2015 27

 Multipath
❑ Signals from satellites at low elevations – more
susceptible to multipath
❑ Code ranges – highly affected compared to
carrier phases
❑ Severe case of multipath – loss of lock
❑ With new receiver and antenna design, and
sound pre-survey planning – minimize possible
causes of multipath
❑ Averaging signals over a period of time – also
help reduce the effect of multipath
28
 Multipath
❑ Several methods – to reduce and estimate
multipath effects:
❖ Antenna based mitigation
❖ Improved receiver technology
❖ Signal and data processing
❑ Antennal based solution
❖ Improving antenna gain pattern by choke rings,
creating special antenna designs and arrays
❖ Elimination of multipath signals – possible by
selecting and antenna that takes advantage of single
polarization
29
 Multipath
❑ Antennal based solution
❖ Transmitted signals – right-handed circularly
polarized, reflected signals – left-handed polarized
❖ Choke ring design – optimal multipath rejection for
both L1 and L2 band
❖ Radio frequency absorbent antenna ground planes
❑ Improved receiver technology
❖ Narrow correlator spacing
❖ Enhancing the strobe correlator multipath rejection –
error reduction up to 90 %

30
 Multipath
❑ Antennal based solution
❖ Transmitted signals – right-handed circularly
polarized, reflected signals – left-handed polarized
❖ Choke ring design – optimal multipath rejection for
both L1 and L2 band
❖ Radio frequency absorbent antenna ground planes
❑ Improved receiver technology
❖ Narrow correlator spacing
❖ Enhancing the strobe correlator multipath rejection –
error reduction up to 90 %

31
 Multipath
❑ Signal and Data Processing
❖ Exploring the signal-to-noise ratio (SNR)
❖ Using multiple reference stations
❖ Smoothing carrier phases
❖ Using data combinations
❑ Most effective counter-measure – avoid sites
having potential multipath effect
❑ Placing antenna directly on the ground without
tripod – eliminate indirect paths at least some
❑ Therefore – recommended to avoid reflecting
surfaces in the vicinity 32
Multipath

© Wellenhof et al., 2001

33
 Receiver Noise
❑ Errors associated with the ability of a receiver to
measure a finite time difference
❑ Include errors due to signal processing, clock/signal
synchronization and correlations, receiver resolution,
signal noise and others
❑ High end receivers – less receiver noise
❑ Receiver noise – directly related to thermal noise,
dynamic stress, etc.

34
 Denial of Accuracy and Access
❑ Two techniques of denying access to civilians
❖ Selective Availability (S/A)
❖ Anti-Spoofing (A-S)
❑ Selective Availability (S/A)
❖ In-force since – March 25, 1990
❖ S/A was turned off on May 2, 2000 at about 4:00
Universal Time (UT)
❖ US DoD – reserves the right to implement S/A – for
national security
❖ Purpose of S/A – denial of accuracy to potential
threats
35
 Denial of Accuracy and Access
❑ Selective Availability (S/A)
❖ Implemented by dithering the satellite clock and
manipulating the ephemerides
❖ Velocity error – 0.3 m/s and time 340 s
❖ Accuracy reduced to – 300 m for horizontal and 500
m for height
❖ Direct impact on pseudorange
❖ Fundamental frequency dithered – both code and
carrier pseudorange affected
❖ Effects of S/A – eliminated by differential techniques

36
 Denial of Accuracy and Access
❑ Selective Availability (S/A)

Radial Orbit Error of PRN 21 on day 177 of 1002 with

SA on and on day 184 of 1992 with SA off (Breuer et
al., 1993 in (Hoffmann-Wellenhof et al., 1992)) 37
 Denial of Accuracy and Access
❑ Selective Availability (S/A)
❖ Two different method of degradation
 Delta process
 Epsilon process
❖ Delta process – dithering fundamental frequency of
the satellite clock
❖ Epsilon process – truncation of the orbital
information in the transmitted navigation message –
coordinates of the satellite – not accurate

38
 Denial of Accuracy and Access
❑ Anti-spoofing (A-S)
❖ Spoofing – tampering something intentionally
❖ Spoofer – generates a signal that mimics the GNSS
signal – signals that have the same structure and
frequency
❖ Spoofed signal – controls its transmitted power level
– the receiver will lock into the spoofed signal rather
than the real GNSS signal
❖ Spoofed signal – the message is changed – reveiver
will calculate – incorrect position or time
❖ Spoofing – always an intentional attack
39
 Denial of Accuracy and Access
❑ Anti-spoofing (A-S)
❖ Complexity of spoofing – increased greatly – attacker attempts to
simultaneously spoof more than one GNSS frequency or
constellation
❖ A/S - Unauthorized users – do not get access to P-code
❖ A/S – either turned on or off
❖ When it is on – P-code on L1 and L2 – replaced by Y-code
available only to authorized users
❖ Unauthorized users – do not get access to P-code
❖ Methods of squaring or cross-correlation – make use of the P-
code when A/S in on
❖ Receiver with multi-frequency and multi-constellation – helps to
detect and overcome the possible spoofing attempt
❖ GNSS+INS sensor – also helps in detecting and overcoming
spoofing attempt 40
 User Equivalent Range Error (UERE)
❑ Total error budget – summarized as the user
equivalent range error (UERE) or user range
error (URE)
❑ UERE – transmitted via navigation message
❑ UERE in combination with DoP – allows to
estimate the achievable positioning accuracy
❑ Does not include multipath
❑ UERE – for single frequency – range from 2.2 m
to 14.6 m
❑ For dual frequency – 1.4 to 2.3 m
41
 User Equivalent Range Error (UERE)
❑ If multipath included – for single frequency – 10
– 15 m
❑ Many of these range errors – can be removed or
effectively suppressed – by developing models
❑ Differential techniques – also eliminate many of
these errors

42