Navigation and Guidance: Dr. Shashi Ranjan Kumar

Navigation and Guidance
Dr. Shashi Ranjan Kumar
Assistant Professor
Department of Aerospace Engineering
Indian Institute of Technology Bombay
Powai, Mumbai, 400076 India
Dr. Shashi Ranjan Kumar IITB-AE 410/641 Lecture 3 Navigation and Guidance 1 / 36
Mechanization of Coordinate Frames in INS
Coordinate frame for computation of velocity, position and attitude by

onboard computer
Selection of coordinate frame
Mission requirements
Worldwide navigation capability
Onboard computation complexity
Interface with other avionics subsystems
INS configurations
Space-stabilized mechanization
Strapdown (gimablless) mechanization
Local-level mechanization
⇒ North-slaved
⇒ Unipolar
⇒ Free azimuth
⇒ Wander azimuth
Local Level Mechanization
Local-level systems are normally expressed in the navigation frame and

measured with respect to an ECI frame.
Distinction between the various local-level mechanization categories is in
azimuth torque rate.
These systems maintain two of accelerometer input axes in horizontal plane.
Earth rotation rate vector

Ωie = Ωie cos φ j + Ωie sin φ k

Angular velocity of the geographic frame in inertial space is the sum of the
angular velocity ρ of the geographic frame w.r.t. the rotating earth plus the
angular velocity Ωie of the earth w.r.t. inertial space.
ωE =ρE
ωN =ρN + ΩN = ρN + Ωie cos φ
ωZ =ρZ + ΩZ = ρZ + Ωie sin φ
As the vehicle moves over the surface of the earth, the platform coordinate
system diverges by wander angle α from the geographic or navigation frame.
The divergence rate α̇ is due to the vertical components of the earth’s
rotation and vehicle motion.
Total angular velocity of the geographic coordinate system with respect to
inertial space
ω gt = −φ̇i + (λ̇ + Ωie ) cos φj + (λ̇ + Ωie ) sin φ k

| {z }
ωzg
Rate of change for α is defined as
α̇ = ωzp − ωzg
where ωzp describes motion of the platform relative to navigation frame.

Input torque applied to platform results into azimuth gyroscope commanded
precession rate ωzc , equivalent to ωzp .
North-Slaved System
Vertical platform axis of this system is torqued to maintain alignment with

the geographic (navigation) axes.
Platform must stay aligned with the geographic axes.
In order for this to be true, the wander angle α must be zero.
As the vehicle moves over surface of the Earth, no rotation out of the
geographic axes.
α̇ = 0 = ωzp − ωzg
As ωzc = ωzp , commanded rate
ωzc = ωzg = Ωie sin φ + λ̇ sin φ = (Ωie + λ̇) sin φ
Unipolar System
Platform vertical axis is torqued to maintain wander angle equal to longitude

angle ⇒ α = λ
Platform must be torqued
⇒ To cancel the precession caused by the earth rate and vehicle rate
⇒ To maintain the wander angle equal to the longitude

α̇ = ωzc − ωzg = ±λ̇

Commanded rate

ωzc = ωzg + α̇ = Ωie sin φ + λ̇ sin φ ± λ̇ = (Ωie + λ̇) sin φ ± λ̇

Double sign indicates that the direction of the wander angle is reversed with
respect to the longitude rate when crossing the equator.
Free-Azimuth System
Vertical platform axis in this system is not torqued.

It eliminates the torquing error associated with vertical gyroscope.
ωzc = 0
Platform axes then diverge in azimuth from the geographic axes

α̇ = ωzc − ωzg = −Ωie sin φ − λ̇ sin φ = −(Ωie + λ̇) sin φ

Non-torquing of vertical gyroscope can be considered as an advantage.
Normally vertical gyroscope exhibits poorer drift rate characteristics than
either of the horizontal gyroscopes.
Free-azimuth coordinates would be used in aircraft INS where high platform
angular rates about the vertical axis are required.
Wander-Azimuth System
Vertical platform axis is torqued to compensate only for the vertical

component of the earth rate.
ωzc = Ωie sin φ
It does not cancel the vertical component of the aircraft transport rate.
No attempt to maintain the level axes in a preferred azimuth direction.
Level axes are allowed to rotate freely about the vertical axis.

α̇ = ωzc − ωzg = Ωie sin φ − Ωie sin φ − λ̇ sin φ = −λ̇ sin φ

Inertial platform is initially aligned by first driving it with torquers until the
outputs of the two-level accelerometers are zero.
System ωc α̇
North-slaved (Ωie + λ̇) sin φ 0
Unipolar (Ωie + λ̇) sin φ ± λ̇ ±λ̇
Free azimuth 0 −(Ωie + λ̇) sin φ
Wander azimuth Ωie sin φ −λ̇ sin φ
Table: Local level systems
North-slaved: Latitude and longitude coordinates can be read directly from

the platform azimuth gimbal
Unipolar: Platform azimuth is maintained equal to longitude; can obtain
azimuth and longitude with a single computation
Free azimuth: Needs no accurate torque electronics for the vertical or
z-gyroscope
Wander azimuth: No singularity at the poles. Also, the coordinate
transformations are simplified considerably
Space stable configuration:

A constant orientation with respect to the inertial space
Platform frame coincides with the inertial frame
Sensing of accelerations in inertial reference as primary inputs
Relative velocity and positions relative to the center of the Earth in Cartesian
coordinates: Integration of acceleration twice in inertial frame of reference
Standard position outputs (latitude and longitude): Derived using inertial
Cartesian components.
Strapdown configuration:
Inertial sensors are mounted directly on the host vehicle.
Platform frame coincides with the body frame
Accelerometers rotate (w.r.t. inertial space) in the same way as the vehicle.
Transformation from sensor axes to inertial reference frame is computed rather
than mechanized.
Platform or Instrument Misalignment
Platform or Instrument Misalignment: Manufacturing imperfections

Sensitive axes of the accelerometer triad not perfectly aligned with the
platform axes
Causes the accelerometers to sense a specific force component due to gravity
Platform or Instrument Misalignment
Unit vectors of platform coordinate frame: 1xp , 1yp , 1zp

Unit vectors of accelerometer triad: 1xa , 1ya , 1za
As misalignment angles are usually small, sin θ ≈ θ and cos θ ≈ 1.
Transformation matrix between these two frames
    
1xa 1 ψ −φ 1xp
 1ya  =  −ψ 1 θ   1yp 
1za φ −θ 1 1zp
1a = C pa 1p
Observations:
Main diagonal elements of the misalignment orientation matrix are near unity.
Off-diagonal elements of misalignment orientation matrix are small as
compared to unity.
Two coordinate frames are in near coincidence.
Sensor
Sensor: A device which provides a usable output in response to a specified

measurand. (American National Standards Institute)
A sensor acquires a physical quantity and converts it into a signal suitable for
processing (e.g. optical, electrical, mechanical).
Active element of a sensor: Transducer which converts one form of energy to
another.
When input is a physical quantity and output electrical ⇒ Sensor
When input is electrical and output a physical quantity ⇒ Actuator
Why do we need sensors ?
Inertial Sensor
Inertial sensor: Sensors based on inertia and relevant measuring principles

Gyroscope
Accelerometer
Gyroscope: Inertial properties of a wheel or rotor spinning at high speed
To sense the angle turned through by a vehicle (displacement gyroscopes)
To sense its angular rate of turn about some defined axis (rate gyroscopes)
Development of mechanical gyroscope: Prof. C S . Draper and his
co-workers, at the Massachusetts Institute of Technology.
Precision devices with error rates of less than 0.001◦ /h, to less accurate
sensors with error rates of tens of degrees per hour.
Capability to measure angular rates up to about 500◦ /s
Fundamentals for gyroscope
Gyroscopic inertia
Angular momentum
Precession
Sensor
Gyroscope
Gyroscopic inertia: Defines a direction that remains fixed in inertial space

The establishment of a fixed direction enables rotation to be detected, by
making reference to this fixed direction.
Orientation of the case of instrument with respect to direction of spin axis
may be measured with angle pick-off devices mounted on the gimbals.
Angular momentum: Angular momentum H of a rotating body
H = Iωs
Distribution of mass on a rotor is important.

Angular momentum should be very high, so that the undesired torques that
can act on a rotor and cause errors are virtually insignificant.
Drift: Any undesired movement of the direction of the spin axis.
How can we maximize the angular momentum with given angular velocity,
mass, and size of rotor?
Gyroscope
One way to for produce a very high angular momentum is to have the
majority of the mass of the rotor at its edge.
This is due to the dependence of the moment of inertia on the square of the
distance of its mass element from the centre of rotation.
Is the high angular momentum always beneficial?
A very high angular momentum results in negligible drift, but there could be
some considerable penalties.
Gyroscope would almost certainly be relatively large and heavy, and it may
take many seconds, if not minutes, for the rotor to reach its operating speed.
In a strapdown mode, the associated control system may not be capable of
recording, or ’capturing’, angular rates beyond a few tens of deg/s.
Compromise while selecting a gyroscope for a given application.
Gyroscope
Spin is the rotation of the gyroscope

rotor relative to the gimbal.
Precession is the rotation of the
gimbal, relative to inertial space.
For a freely spinning body, such as
the Earth, there is not a material
frame with spin bearings.
The precession must be considered
to be that of the axis system which
an imaginary gimbal would have one
axis through the north and south
poles, and two mutually orthogonal
in the plane of the Equator.
Gyroscope
Consider spinning of disc about ss0 .

If the disc is acted upon by a couple, that is, a torque, the torque being about
the axis tt0 , the spin axis of the disc will be forced to turn about the axis pp0 .
This turning is called as the precession.
The precession axis, pp0 , is orthogonal to the torque axis, tt0 .
Suppose that the disc is rigid with all the mass in the rim and that the rim
has a peripheral speed u.
Consider an element of mass at the highest point, P1 .
The instantaneous velocity of the mass is changed by adding the velocity w
in the same sense as the couple F F .
The resultant velocity, v, is now in a different direction.
Other elements of the rim change their velocities in proportion to their
distance from the axis tt0 .
Gyroscope
After the disc has spun through 90◦ , the element of mass arrives at the point
P2 , which is not in the expected line tt0 , but in a plane which has precessed
about the axis pp0 .
Gyroscope: Mathematical Description of Precession
A heavy spinning disc with angular momentum H defined by vector OA, Ha

with unit vector a.
Newton’s law: The angular momentum vector H remains constant unless
the some external torque acts on the disc.
Suppose a torque T is applied to the disc and it causes a precession at rate

ω = ωc.
This axis lie in the plane of the disc and may be taken to be normal to the
plane of the paper.
Over time duration δt, the disc will precess through an angle ωδt about c.
Angular momentum vector will change to OB, that is, (H + δH)b where
b = a + ωδt(c × a).
Change in angular momentum
δH =(H + δH)b − Ha = H(b − a) + δHb = Hωδt(c × a) + δHb
Rate of change of angular momentum

dH δH dH dH
= lim = Hω(c × a) + b=ω×H + b
dt δt→0 δt dt dt
dH
T =ω × H + b
dt
Component of the torque which is along the spin axis gives rise to an
acceleration in the spin rate.
Component normal to the spin axis gives rise to a precession ω which is
normal to both the torque and the spin axes.
The direction of the precession is such as to try to align the spin axis with
the torque axis.
On neglecting component along spin axis, we obtain “Law of Gyroscopes” as

T = ω × H ⇒ T = ωH
Measurement of angular rotation of rotation rate:
Change in angle of gimbals
Measurement of torque to keep rotor aligned with direction defined by
instrument case
Gyroscope
Difficult to measure angular displacements accurately, need expensive

equipments
Easy to measure fixed defined position or zero deflection or null position
If the spin axis of a rotor is made to precess back to the ‘null’ position by the
application of a suitable torque, a very accurate angular measurement is
possible.
Requirements: Necessary torque to null the deflection can be generated and
measured.
Even most accurate of gyroscopes will appear to drift.
This is because the angular momentum vector is fixed with respect to space
axes, not the co-ordinate system defined by the Earth.
It is sometime necessary to apply corrective torques to precess the gyroscope
if it is to be used as an Earth reference.
Sensor Errors Types
Fixed or repeatable terms:

A bias component which is predictable
Always present whenever the sensor is switched on
Can therefore be corrected
Temperature induced variations:
A temperature-dependent bias component
Can be corrected with suitable calibration
Switch-on to switch-on variations:
A random bias which varies from gyroscope switch-on to switch-on
Constant for any one run
In-run variations:
An in-run random bias which varies throughout a run
Precise form of this error varies from one type of sensor to another
First two may be corrected to large extent, but later two types of errors
persists.
Ring Laser Gyro
Optical sensor: Sensors using properties of EM radiation to sense rotation.

Ring Laser Gyro (RLG): Principles of general relativity
A resonant optical cavity containing two oppositely directed traveling light
waves generated by stimulated emission of radiation
An inherent capability to operate in the strapdown mode
Scale factor linearity over the full dynamic range
Easy interface with digital systems
Detection of rotation with light was demonstrated by Sagnac in 1913.
Principle: Two light waves acquired a phase difference by propagating in
opposite directions around a loop interferometer
These beams travel the same path in opposite directions around a closed ring.
Solid-state sensor with no moving parts.
It detects and measures differential angular rotations by measuring the
frequency difference between the two contrarotating beams.
Passive Sagnac Interferometer
Consider a circular interferometer.

Assume the light beam is splited into two beams rotating in opposite
directions.
After one rotation they combine at the beamsplitter.
In absence of rotation, transit time taken by light to complete rotation
2πR
t=
c
where, c is the speed of light, and R is the radius of circular path.
If the interferometer is rotated with constant angular velocity Ω then the
travel time for both the beams will be different.
Note that the beamsplitter is moved to new location at B.
With respect to inertial space, the light moving in the direction of rotation
must travel a longer distance than the light traveling in the opposite direction.
Let X be the inertial space distance between points A and B.

Positive (+) and negative (−) signs refer to the beam traveling in the
direction of rotation and opposite to the direction of rotation, respectively.
Total closed-path transit for the light
ct± =2πR ± X± , X± = RΩt±
On solving above equations,

2πR RΩt± 2πR
t± = ± ⇒ t± =
c c c ∓ RΩ
Transit time ∆t
2πR 2πR
∆t =t+ − t− = −
c − RΩ c + RΩ
4πR2 Ω

2RΩ
=2πR 2 =
c − R 2 Ω2 c2 − R2 Ω2
Transit time ∆t

4πR RΩ
4πR2 Ω c c
∆t = = 2
c2 − R 2 Ω2

RΩ
1−
c
On neglecting smaller terms, we get equation called as “Sagnac effect”
4πR2 Ω 4πΩR2

4πR RΩ 4AΩ
∆t = 2 2 2
= = 2
= 2
c −R Ω c c c c
where, A = 4πR2 is the area of circular optical path.
Optical path difference ∆L
4cπΩR2 4ΩπR2 4AΩ

∆L =c∆t = = =
c2 c c
This result holds in general for any geometric closed path.
Issue: The path difference is small even with a large area.
If Ω is very low then it is difficult to measure these angular rates as required
closed area is very large.
Ratio of total enclosed area to the wavelength must be large to sense low
angular rate.
Lack of sensitivity because the path difference for light traveling in the two
directions is much less than wavelength.
Active Ring Laser Interferometer
Improvement of sensitivity
By replacing the beamsplitter with a mirror
Form a resonant circuital optical cavity supporting traveling-wave modes for
the counterrotating beams.
These modes could be made self-sustaining by placing the lasing medium in
the cavity.
Laser frequency is dependent on the cavity length.
Two oppositely directed traveling waves oscillate independently, each with its
own frequency and amplitude.
Fractional difference between these two frequencies corresponds to the
fractional difference in optical path lengths traveled by each wave and,
therefore, is proportional to the angular velocity.
To sustain oscillation, there must be enough gain in the medium to overcome

losses in the cavity.
Optical length of beam also need to satisfy
N λ± = L±
where, L± is the optical length of each beam, λ± is the wavelength, and N

is large integer (105 to 106 ).
Cavity geometry determines the wavelengths of a given mode.
Fractional frequency shift equals the fractional path length
∆ν ∆L
= (Proof?)
ν L
As λ = c/ν, we have beat frequency ∆t given by

∆Lν 4AΩ c 4A
∆ν = = = Ω
L c λL Lλ
Ideal RLG equation

4A
∆ν = Ω
Lλ
| {z }
Geometric or ideal scale factor S
where, λ and L are the wavelength of laser light and optical path length or
cavity length, respectively.
Laser gyro is rate integrating gyro as it gives N counts when turned through
angle θ.
On integration of ideal RLG equation
Z t2 Z t2
∆νdt =S Ωdt ⇒ N = Sθ
t1 t1
where, N is total phase shift or beats counted during measurement time and
θ is total angle of rotation.
Text/References
Reference
1 G. M. Siouris, Aerospace Avionics Systems: A Modern Synthesis, Academic
Press, Inc. 1993.
2 D. H. Titterton and J. L. Weston, Strapdown Inertial Navigation Technology,
Progress in Astronautics and Aeronautics, Vol. 207, ed. 2, ch. 4.
Thank you for your attention !!!

Navigation and Guidance: Dr. Shashi Ranjan Kumar

Încărcat de

Informații document

Descriere originală:

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

Navigation and Guidance: Dr. Shashi Ranjan Kumar

Încărcat de

Drepturi de autor:

Formate disponibile

Navigation and Guidance

Dr. Shashi Ranjan Kumar

Coordinate frame for computation of velocity, position and attitude by

Local-level systems are normally expressed in the navigation frame and

ω gt = −φ̇i + (λ̇ + Ωie ) cos φj + (λ̇ + Ωie ) sin φ k

Rate of change for α is defined as

where ωzp describes motion of the platform relative to navigation frame.

Vertical platform axis of this system is torqued to maintain alignment with

As ωzc = ωzp , commanded rate

ωzc = ωzg = Ωie sin φ + λ̇ sin φ = (Ωie + λ̇) sin φ

Platform vertical axis is torqued to maintain wander angle equal to longitude

Vertical platform axis in this system is not torqued.

Platform axes then diverge in azimuth from the geographic axes

Vertical platform axis is torqued to compensate only for the vertical

ωzc = Ωie sin φ

North-slaved: Latitude and longitude coordinates can be read directly from

Space stable configuration:

Platform or Instrument Misalignment: Manufacturing imperfections

Unit vectors of platform coordinate frame: 1xp , 1yp , 1zp

Sensor: A device which provides a usable output in response to a specified

Inertial sensor: Sensors based on inertia and relevant measuring principles

Gyroscopic inertia: Defines a direction that remains fixed in inertial space

Distribution of mass on a rotor is important.

Spin is the rotation of the gyroscope

Consider spinning of disc about ss0 .

A heavy spinning disc with angular momentum H defined by vector OA, Ha

Suppose a torque T is applied to the disc and it causes a precession at rate

δH =(H + δH)b − Ha = H(b − a) + δHb = Hωδt(c × a) + δHb

Rate of change of angular momentum

Difficult to measure angular displacements accurately, need expensive

Fixed or repeatable terms:

Optical sensor: Sensors using properties of EM radiation to sense rotation.

Consider a circular interferometer.

Let X be the inertial space distance between points A and B.

ct± =2πR ± X± , X± = RΩt±

On solving above equations,

On neglecting smaller terms, we get equation called as “Sagnac effect”

where, A = 4πR2 is the area of circular optical path.

Optical path difference ∆L

4cπΩR2 4ΩπR2 4AΩ

To sustain oscillation, there must be enough gain in the medium to overcome

where, L± is the optical length of each beam, λ± is the wavelength, and N

Ideal RLG equation

Thank you for your attention !!!

S-ar putea să vă placă și