ashby gps relativity
TRANSCRIPT
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 1/35
July 20, 2006 AAPT workshop 1
Relativity in the Global Positioning System
Neil Ashby
Department of Physics,UCB 390
University of Colorado, Boulder, CO 80309-00390
NIST Affiliate
Email: [email protected]
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 2/35
July 20, 2006 AAPT workshop 2
Fundamental Principles
• Principle of Inertia– The laws of physics are the same in all inertial frames of reference.
• Constancy of the speed of light– The speed of light, c, is a constant independent of the motion of the
source (or of the observer); (it is c in all inertial frames).
• Principle of Equivalence (“weak form”)– Over a small region of space and time, the fictitious gravitational field
induced by acceleration cannot be distinguished from a realgravitational field due to mass.
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 3/35
July 20, 2006 AAPT workshop 3
Changes in point of view (reference frames)
In studying relativity, one must be willing to adopt different points of view--
that is, different reference frames. The physical phenomena
don’t change, but our description of them does change.
Example 1: the gulf stream curves toward the east as it flows north. For an
observer fixed on the rotating earth, this is due to the “Coriolis force.”
To an observer in a local, freely falling non-rotating frame attachedto earth’s center, this is due to conservation of angular momentum.
Example 2: A pendulum in an accelerating car points backwards. From the
point of view of someone in the car, the force of gravity points slightly
downwards and slightly backwards.
From the point of view of someone on the ground, the pendulum bob
is accelerated forwards by a component of tension in the string that
points slightly forwards and upwards.
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 4/35
July 20, 2006 AAPT workshop 4
Question.
In an accelerating vehicle, if up is indicated by the direction of the string
that holds down a helium balloon, what direction is “up”?
Up?Up?
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 5/35
July 20, 2006 AAPT workshop 5
Relativity of Simultaneity
To an observer on the ground, let two lightning
strokes at the front and back of the train
be simultaneous.
The “moving” observer at the train’s midpoint findsthe event at front occurs first.
2'
vxt t
c≈ −
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 6/35
July 20, 2006 AAPT workshop 6
Notation--”Lab” and “Moving” Frames
*
Event
In Lab Frame: { t, x ,y, z }
In Moving Frame: { t’, x’, y’, z’ }
x,x’
y y’
z z’
Origins pass by each other at t = t’ = 0.
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 7/35
July 20, 2006 AAPT workshop 7
c
Breakdown of simultaneity
In a given inertial frame, it’s OK to take differences of velocities and obtain a
velocity difference greater than c.
Example: Let a rod of length L=x move in the positive x-direction with speed v.
Light emitted at t=0 from the left end of the rod travels to the right end.v
vc
In Lab, arrival time at the rod end:
2;
( ) (1 / )
L L L vLt
c v c v c c c= = +
− −≃
L
In the moving frame:
2 2' ; '
L vL vxt t t t
c c c= = − = −
event at t=t’=0
signal propagating with speed c
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 8/35
July 20, 2006 AAPT workshop 8
Breakdown of simultaneity
x,x’
y y’
z z’
*
Event
2'
vxt t
c−≃
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 9/35
July 20, 2006 AAPT workshop 9
Relation Between Doppler Effect
and Relativity of Simultaneity
0 2 3 ....... ... x nλ λ λ λ =
2'
vnt
c
λ = −
0t =
Wavefronts are marked
simultaneously by the
“non-moving” observer:
Moving observer says the wavefronts are marked differently:
So the wavefront at ( 1) x nλ = =
before it gets into the right position to be marked at t’ = 0. To the moving observer,
the wavelength is:
2
v vc
c c
λ λ =
' (1 ) .v v
c cλ λ λ λ = + = +
c
needs to move an additional distance
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 10/35
July 20, 2006 AAPT workshop 10
Doppler frequency shift
, ln ln ln .c f c f λ λ = = +
So when lambda increases, the frequency decreases. Taking the
differential of the logarithm function gives
. f v
f c
λ
λ
∆ ∆= − = −
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 11/35
July 20, 2006 AAPT workshop 11
Equivalence Principle and Gravitational Frequency Shifts
Over a small region of space and time,a fictitious gravity field induced by
acceleration cannot be distinguished
From a gravity field produced by mass.
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 12/35
July 20, 2006 AAPT workshop 12
Gravitational Frequency Shift
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 13/35
July 20, 2006 AAPT workshop 13
2 2
2
1
2
/
;
.
(1 )2
t L c
gL
v gt c
f v gL
f c c c
J GM GM r a f
f c
=
= =
∆ ∆Φ= − = − = −
− − − +
∆ = −
Gravitational Frequency Shift
The situation in the rocket is static. The fractional
frequency difference between the clocks is
2.
f
f c
∆ ∆Φ= +
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 14/35
July 20, 2006 AAPT workshop 14
Frequency shifts due to Gravitational Potential Differences
Let be the gravitational potential on earth’s geoid--at mean sea level,
and r be the radius of a GPS satellite.0Φ
0 E GM
r ∆Φ = − − Φ
0Φ includes contributions from earth’s oblateness as well as its mass:
0 2
1
1(1 ),
2
GM J
a
Φ = − +
where: a1 is the equatorial radius of the earth;
2 .001086 J ≃ is earth’s quadrupole moment coefficient.
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 15/35
July 20, 2006 AAPT workshop 15
How big are gravitational frequency shifts in the GPS?
2
1
1( )
f GM GM
f c a a ∆
= − − −
To get a rough estimate, assume the satellite orbit is circular and
the reference clock is on earth’s equator.
where
105 10 ; ( 13km navigation error per day) f
f
−∆× ≈≃
14 3 2
1
3.986 10 m /s ;
26,562km;
6,378km
GM
a
a
= ×
=
=
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 16/35
July 20, 2006 AAPT workshop 16
Best IIR Rbs
Navs 51, 54
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 17/35
July 20, 2006 AAPT workshop 17
Question:
If a clock makes an error in one day of 1 part in 10^14, how far
would light travel in this amount of time?
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 18/35
July 20, 2006 AAPT workshop 18
Question:
If a clock makes an error in one day of 1 part in 10^14, how far
would light travel in this amount of time?
In one day, the error is:
14 10
10 86400sec 8.64 10 sec.− −
× = ×
In this amount of time, light travels a distance
10 10(8.64 10 sec) 299792458m/sec (8.64 10 sec)
0.26 m
d c− −
= × × = × ×
=
Answer:
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 19/35
July 20, 2006 AAPT workshop 19
Constancy of c
( ) ,
1,2,3,4
j jc t t
j
= −
=
r - r
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 20/35
July 20, 2006 AAPT workshop 20
Sample data
20959777.407-8705178.66813799439.29437239.929 346 353 94
3486595.54615908390.24520982631.27037239.
925 307 870 03
-4436309.87516360459.35820450127.56637239.
920 713 391 82
13246718.72118997823.89513004579.64237239.924 422 365 61
Transmitter
Position z i (m)
Transmitter
Position y i (m)
Transmitter
Position x i (m)
Transmission
Epoch t i (s)SV #
Question: Where is the receiver and what is the time at the receiver?
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 21/35
July 20, 2006 AAPT workshop 21
Reciprocity
( ) j jc t t = −r - r
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 22/35
July 20, 2006 AAPT workshop 22
The constancy of the speed of light implies time dilation
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 23/35
July 20, 2006 AAPT workshop 23
Einstein’s Light Clock
2 2c v−
2 2
2 2
2 2
/ / ;
1 /
' / 1 /
L ct L c v
v c
t L c v c t
= − =
−
= = −
This is just Euclidean geometry and the constancy of c.
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 24/35
July 20, 2006 AAPT workshop 24
How Big is Time Dilation in the GPS?
22 2
2
211
2
11 / 1 ;2
4000m/s;1
3.5 102
vv cc
vv
c
−
− ≈ −
=
− = − ×
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 25/35
July 20, 2006 AAPT workshop 25
Accounting For Relativistic Effects
Example: Time Dilation:
( )
2 2
1/ 22 2
2
2
2
2
path
1 / ;
1 /
1
1 .2
11
2
d v c dt
dt v c d
v
d c
vt d
c
τ
τ
τ
τ
−
= −
= −
+
∆ = +
∫
≃
Elapsed Coordinate time:
Observed Proper Time
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 26/35
July 20, 2006 AAPT workshop 26
GPS Satellite in a circular earth-bound orbit
3;
GmM m
r
− =r rɺɺ
Newton’s law of motion: Force toward the earth is gravitational.
The mass of the satellite cancels out (a consequence of the
Principle of Equivalence). The orbit is nearly a Kepler ellipse.
Solution of the above equation shows that the energy is constant:
21
2 2
GM GM v
r a− = −
where a is the semimajor axis.
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 27/35
July 20, 2006 AAPT workshop 27
Frequency shift of GPS satellite clocks
relative to reference clock on equator
( )22
10
2 2 2
1 1
2 2
E a f v
f c c c
ω Φ − Φ∆= − − −
( )2
2 12
1
10
2
31 1 1 12 (1 / 2)
2 22 1 1
4.4647 10 .
E E E E
E
GM GM f GM J a
f c a r a aGM
c a r
ω
−
∆ = − − − + + +
= − − + ×
Putting in everything that is known about these quantities and addingand subtracting some terms,
The first term depends on orbital eccentricity and gives rise to a periodic timeerror that is ( E is the eccentric anomaly)
2
2sin( ) .
rel E
et GM a E const
c∆ = − +
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 28/35
July 20, 2006 AAPT workshop 28
Clocks on earth’s geoid beat at equal rates
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 29/35
July 20, 2006 AAPT workshop 29
Orbit adjustments in GPS
2 2
3 3;
2 2
E E GM GM f a
f c a c a a
δ δ ∆ = ∆ − =
For an increase in altitude of 20 km, the change in frequency is
131.88 10 . f f
δ −∆ ×≃
This effect is now understood and the clock frequency is adjustedwhen the orbit is changed. (Back-of-the-envelope calculation!)
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 30/35
July 20, 2006 AAPT workshop 30
END
Further development--introduce the metric
2 2 2 2 2
2
2(1 )( ) ;ds c dt dx dy dz
c
Φ= − + + + +
Discuss proper time, coordinate time; show how the effects can be
obtained from the metric…..
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 31/35
July 20, 2006 AAPT workshop 31
Sagnac effect
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 32/35
July 20, 2006 AAPT workshop 32
Sagnac Effect on Synchronization
in a Rotating System
Sagnac Correction =2
2 z A
c
ω
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 33/35
July 20, 2006 AAPT workshop 33
Principle of Equivalence
V
V+∆∆∆∆V
Acceleration
Local Inertial
Frame
r
2
2 2
22
2
2 2 2
Induced potential difference/c ;
Gravitational potential difference/c ;
1Net potential difference/c = 0;
c Tc
c
c T c c
∆= =
∆
∇Φ=
+ ∇Φ ∆ ∇Φ= + =
∆
A r V r
r
A r r V r r
i i
i
i i i i
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 34/35
July 20, 2006 AAPT workshop 34
Common-view time transfer
8/22/2019 Ashby GPS Relativity
http://slidepdf.com/reader/full/ashby-gps-relativity 35/35
July 20, 2006 AAPT workshop 35
Why are atomic clocks needed?
To reduce the effect of clock error to < 2 meters,
the clock error must be less than 2/c = 6.7 x 10-9 sec.
Half a day = 43200 seconds, so the fractional clock
error must be less than:
(2 m)/(43200 s x c) = 1.5 x 10-13.
Only atomic clocks can achieve such stability.