Control For Control Sustem

Chapter 12 Motion Control Systems
12.1 Open-Loop Stability and Maneuverability

12.2 PID Control and Accelera>on Feedback
12.3 Control Alloca>on

Chapter 12 covers state-of-the-art PID

control methods for:

Setpoint regula>on
Trajectory-tracking control
Path-following control

of marine craK. This includes autopilot design, sta>onkeeping, posi>on mooring systems,
cross-tracking control systems and LOS control systems. In addi>on to this, control
alloca>on methods are discussed.

Advanced methods such as linear-quadra>c op>mal control, sliding mode control, state
feedback lineariza>on and integrator backstepping are discussed in Chapter 13.
1

Lecture Notes TTK 4190 Guidance and Control of Vehicles (T. I. Fossen)

Chapter 12 Motion Control Systems
Control is the ac)on of determining the necessary control forces and moments to be
provided by the cra6 in order to sa)sfy a certain control objec)ve.

Control Alloca+on: Distribu>on of generalized control forces to the actuators.
wind
loads
wind

feedforward

pilot yaw angle
input and rate
reference control control compass
model system allocation and gyro
observer and
wave filter
2


12.1 Open-Loop Stability and
Maneuverability
Stability Maneuverability
Copyright Bjarne Stenberg

Stability of the uncontrolled ship can be dened as the ability to return to an equilibrium point
aKer a disturbance, without any correc>ve ac>on of the rudder.

Maneuverability, on the other hand, is dened as the capability of the ship to carry out
specic maneuvers.

Excessive stability implies that the control eort will be excessive in a maneuvering situa>on
whereas a marginally stable ship is easy to maneuver. Thus, a compromise between stability
and maneuverability must be made.
3


12.1.1 Straight-Line, Directional and
Positional Motion Stability
For ships it is common to dis>nguish between three types of stability:
Straight-line stability
Direc>onal stability
Posi>onal mo>on stability
Consider the following test system:
x! " u cos ! ! v sin ! " u 0 cos ! # cruise speed u0

y! " u sin ! # v cos ! " u 0 sin ! #
!! " r

#
Nomoto model Tr! # r " K" # w #
PD controller ! ! !Kp !" ! " d" ! Kd r
Resul>ng ! " !1 " KKd "!# " KKp ! $ KKp ! d " w #

T!
test system
% #
! " d!# " k! $ f!t"
m! #
4


Disturbance:
impulse at t = 100 s
Test system:
! " d!# " k! $ f!t"

m! f!t" ! k! d " w
!1,2 !
!d" d 2!4km
, "n !

,
k

#
! d 1
2m m 2 km
Matlab MSS toolbox script: Stabdemo.m
5
Copyright Bjarne Stenberg/SINTEF

Straight-Line Stability: Consider an uncontrolled ship (Kp = Kd = 0) moving on a straight-line
path. If the new path is straight aKer a disturbance w in yaw the ship is said to have
straight-line stability.
XY-Plot: Straight-line stability

200
The direc>on of the new path
150
will usually dier from the ini>al
100
path because no restoring
50
forces are present (k = 0). This
corresponds to:
0
-50

d ! !1 " 0
!1 ! ! m
-100
0 500 1000 1500 2000 2500 3000
T r (deg/s) (deg)
!2 ! 0 0.06 6
0.05 5
0.04 4
The requirement T > 0 implies 0.03 3
straight-line stability for the 0.02 2
uncontrolled ship ( = 0) 0.01 1
0
0 200 400 600
0
0 200 400 600
sec sec
6


Direc+onal Stability (Stability on Course): Direc>onal stability requires the nal path to be
parallel to the ini>al path which is obtained for Kp > 0. Addi>onal damping is added through
Kd > 0, that is, PD-control.

The ship is said to be direc>onally stable if both eigenvalues have nega>ve real parts, that is:
XY-Plot: Directional stability (critical damped)
10

Re!!1,2 " ! 0 5
Two types of direc>onal stability are

0

observed:
-5
0 500 1000 1500 2000 2500
(deg)
3000
1. No Oscilla+ons: (d - 4km 0): This r (deg/s)
implies that both eigenvalues are 0.06 0.7
nega>ve and real-i.e. 1 such that:

0.05 0.6
0.04 0.5
2
0.03 0.4
!d " d ! 4km 0.02 0.3

!1,2 !
2m 0.01 0.2
0 0.1
! !" " "2 ! 1 #n # 0 -0.01

0 200 400 600
0
0 200 400 600
sec
7


2. Damped Oscillator (d - 4km < 0):
This corresponds to two imaginary eigenvalues with nega>ve real parts ( < 1),
that is:
!d " j 4km ! dm XY-Plot: Directional stability (underdamped)

!1,2 ! 10
2m
! !" " j 1 ! " 2 # n 5

0

No>ce the oscilla>ons in both
posi>ons and yaw angle. -5
0 500 1000 1500 2000 2500 3000
Direc>onal stability requires
feedback control since there are 0.06
r (deg/s)
1.5
(deg)
no restoring forces in yaw.

0.04 1
However, in heave, roll and pitch
where metacentric restoring forces 0.02 0.5
are present (k > 0), no feedback is 0 0
required to damp out the

oscilla>ons. -0.02
0 200 400 600
-0.5
0 200 400 600
sec sec
8


Posi+onal Mo+on Stability: Posi>onal mo>on stability implies that the ship should return
to its original path aKer a disturbance. This can be achieved by including integral ac>on in
the controller.
A PID controller can be

designed to compensate for
the unknown disturbance

term w while a PD controller
will generally result in a
steady-state oset.
9


12.1.2 Maneuverability
Ship maneuvers can be used to evaluate the robustness, performance and limita>ons
of a ship.
This is usually done by dening a criterion in terms of a maneuvering index or by using
a maneuvering characteris>c.
The Norrbin Measure of Maneuverability

! " !t " !1" Heading change per unit rudder angle in one ship length traveled at U=1 m/s
P! " " !t " !1" t= t(U/L) is the non-dimensional >me
T! !" ! # !$ ! % K ! " !

exp!!t /T " " 1 !
! ! t!
#
!t ! " 2
# O!3"
T! 2!T ! " 2
! ! !t! " " K ! #t ! ! T ! # T ! exp!!!t ! /T ! ""$" ! !t ! "

! ! !t ! " ! ! ! ! t! !t ! " 2 !t ! " 2
" ! !t ! "
! K t "T "T 1" T!
" 2!T ! " 2
# K 2T !
! ! !t ! "1" ! !t ! " 2 K! 1 K!
" ! !t ! "1"
!K 2T !
" 2T ! P! 2 T!
t ! "1
Norrbin concludes that P > 0.3 guarantees a reasonable standard of course-change quality for
most ships, while P > 0.2 seems to be sucient for large oil-tankers.
10


Maneuvering Characteris+cs
A maneuvering characteris>c can be obtained by changing or keeping a predened course
and speed of the ship in a systema>c manner by means of ac>ve controls. A guide for sea
trials describing how these maneuvers should be performed is found in SNAME (1989). The
following ship maneuvers have been proposed by ITTC:

Turning Circle: This trial is mainly used to calculate the ship's steady turning radius and to
check how well the steering machine performs under course-changing maneuvers.
Kempf's Zigzag Maneuver: The zigzag test is a standard maneuver used to compare the
maneuvering proper>es and control characteris>c of a ship with those of other ships.

Experimental results of the test can be used to calculate the K and T values of Nomoto's
1st-order model.
Pull-Out Maneuver: The pull-out maneuver can be used to check whether the ship is
straight-line stable or not.
Dieudonn's Spiral Maneuver: The spiral maneuver is used to check straight-line stability.
The maneuver gives an indica>on of the range of validity of the linear theory.
Bech's Reverse Spiral Maneuver: The reverse spiral maneuver can be used for unstable
ships to produce a nonlinear maneuvering characteris>c. The results from the test
indicate which rudder correc>ons that are required to stabilize an unstable ship.
Stopping Trials: Crash-stops and low-speed stopping trials can be used to determine the
ship's head reach and maneuverability during emergency situa>ons.
11


Turning Circle
The turning circle is probably the oldest maneuvering test. The test can be used as an
indica>on on how well the steering machine and rudder control performs during course-
changing maneuvers. It is also used to calculate standard measures of maneuverability such
as tac>cal diameter, advance and transfer.

Matlab MSS
toolbox script
ExTurnCircle.m
12


The steady turning radius R is perhaps the most interes>ng quan>ty obtained from the
turning trials. Steady-state analysis gives:
!"$0
M!" #N!u0 "! ! b! # ! ! N!1!u 0"b!
!Y v N!!Nv Y !"
r! Y v !Nr !mx g u 0"!Nv !Y r !mu 0"
!
The ship's turning radius R is dened as:

U
R! r where U ! u2 " v 2
This gives the non-dimensional formula for the turning radius:
R Y "v !N "r !m " x "g "!N "v !Y "r !m " " 1

! "!L Y "v N "!!N "v Y "! ! , !"0
13


Example 12.2 (Determina+on of the Nomoto Gain and Time Constants)
The Nomoto gain and >me constants can be computed from a turning test by using
nonlinear least-squares curve gng. For a step input = 0, we get:

Tr! " r # K! r!t" ! exp!!t/T"r!0" " #1 ! exp!!t/T"$K! 0

where K and T are unknowns. The Matlab MSS toolbox script ExKT.m ts this model to a
simulated step response of the model mariner.m which is a nonlinear model of the Mariner
class vessel.

The results for a step d0 = 5 deg and U = 7.7 m/s = 15 knots, are
K = 0.09 s-1 , T = 22.6 s
The Norrbin measure of maneuverability becomes:
P = (K/T ) = 0.87
which guarantees good maneuverability since P > 0.3.
14


% ExKT Script for computation of Nomoto gain and time constants % time-series
% using nonlinear least-squares.The rudder input is 5 deg at t!0. tdata ! xout(:,1);
rdata ! xout(:,2)*180/pi;
N ! 2000; % number of samples
h ! 0.1; % sample time
% nonlinear least-squares parametrization:
% x(1)!1/T and x(2)!K
xout ! zeros(N,2);
x ! zeros(7,1);
x0 ! [0.01 0.1]
delta_R ! 5*(pi/180); % rudder angle step input F ! inline(exp(-tdata*x(1))*0 "...
x(2)*(1-exp(-tdata*x(1)))*5,x,tdata)
for i!1:N, x ! lsqcurvefit(F,x0, tdata, rdata);
xout(i,:) ! [(i-1)*h ,x(3)];
xdot ! mariner(x,delta_R); % nonlinear Mariner model

r!t"
! exp!!t/T"r!0" " #1 ! exp!!t/T"$K!
x ! euler2(xdot,x,h); % Euler integration
end 0
15


Kempf's Zigzag Maneuver
The zigzag test was rst proposed by Kempf (1932).

20o-20o Maneuver
(1st angle = rudder command, 2nd angle = change of heading before the rudder is reversed)

The zigzag >me response is obtained by moving the rudder to 20o starboard from an
ini>ally straight course. The rudder segng is kept constant un>l the heading is changed 20o,

then the rudder is reversed 20o to port. Again, this rudder segng is maintained un>l the
ship's heading has reached 20o in the opposite direc>on. This process con>nues un>l a total
of 5 rudder step responses have been completed.

Standardized by the Interna>onal Towing Tank Conference (ITTC) in 1963.

For larger ships, ITTC has recommended the use of a 10o-10o or a 20o-10o maneuver to
reduce the >me and water space required.
16


20o-20o Maneuver

17


20o-10o Maneuver

18


Pull-Out Maneuver
In 1969 Roy Burcher proposed a simple test procedure to determine whether a ship is
straight-line stable or not. This test is referred to as the pull-out maneuver (12th ITTC ).
The pull-out maneuver involves a pair of maneuvers in which a rudder angle of
approximately 20o is applied and returned to zero aKer steady turning has been
akained.
Both a port and starboard turn must be performed.

During the test the ship's rate of turn must be measured or at least calculated by numerical
deriva>on of the measured compass heading.

Straight-line stable: the rate of turn will decay to the same value for both the starboard
and port turn.

Unstable: the steady rate of turn from the port and starboard turn dier.

19


Pullout maneuver for the Mariner class Pullout maneuver for a tanker. No>ce
vessels. No>ce that the posi>ve and nega>ve that the posi>ve and nega>ve curves do
curves meet for the stable ship. not meet for the unstable ship.
20


Dieudonn's Spiral Maneuver
The direct spiral test was published rst in 1949-1950 by the French scien>st Jean
Dieudonn. An English transla>on was available in 1953.
The direct spiral maneuver is used to check straight-line stability. The maneuver also
gives an indica>on of the degree of stability and validity of linear theory.
To perform the test the ship should ini>ally be held on a straight course. The rudder
angle is then put to 25o starboard and held un>l steady yawing rate is obtained. AKer
this the rudder angle is decreased in steps of 5o and again held un>l constant yawing
rates are obtained for all the rudder angles. The procedure is performed for all rudder

angles between 25o starboard and 25o port. In the range around zero rudder angle the
step of 5o rudder should be reduced to obtain more precise values.

The results are ploked in an r-d diagram.

For straight-line unstable ships it is recommended to use Bech's reverse spiral maneuver
because of the hysteresis at small turning rates.
21


r-d diagram showing the Dieudonne and Bech spirals for both a stable and unstable ship.
No>ce the hysteresis loop in the Dieudonne spiral for the unstable ship
22


Bech's Reverse Spiral Maneuver
For unstable ships within the limits indicated by the pull-out maneuver, Bech's reverse spiral
should be applied. The reverse spiral test was published in 1966.
The ship steering characteris>c is nonlinear outside a limited area. The mean value of the
required rudder deec>on ss to steer the ship at a constant rate of turn rss is a nonlinear
func>on:
! ss ! H B !r ss "
where HB(rss) is a nonlinear func>on describing the maneuvering characteris>c.

This can be understood by considering the nonlinear model:
T1 T2 r! " !T 1 " T 2 "r# " KH B !r" $ K!! " T 3 !# "
If r ! rss is constant in steady-state, that is, r" ! r# ! !# ! 0 directly gives ! ss ! H B !r ss "
The full-scale test is performed by measuring the necessary rudder ac>on required to bring
the ship into a desired rate of turn. For an unstable ship this implies that the rudder angle
will oscillate about a mean rudder angle. The amplitude of the rudder oscilla>ons should
be kept to a minimum. AKer some >me a balance condi>on is reached and both the
mean rudder angle and rate of turn can be calculated.
23


12.2 PID Control and Acceleration
Feedback
This sec>on discusses PID control design for SISO and MIMO mo>on control systems
and it will be shown how PID controllers can be designed to exploit accelera>on
feedback in marine systems.
Accelera>on feedback can be implemented in conjuncture with PID-control
without increasing the demand for control energy.
It is also possible to increase (or even decrease) the mass/iner>a of the closed-loop
system using accelera>on feedback. In addi>on accelera>on feedback can be used
to shape the system iner>a matrix e.g. compensa>on of added mass.

By increasing the mass of the system and thus the iner>a the closed-loop system
will be more robust to environmental disturbances to the price of less
maneuverability.
A mass-damper-spring system will be used to demonstrate the main concept.
Further reading:
K.-P. Lindegaard (2003). Accelera>on Feedback in Dynamic Posi>oning Systems, PhD thesis, Department
of Engineering Cyberne>cs, NTNU
24


12.2.1 Linear Mass-Damper-Spring Systems
mx! " dx# " kx $ 0 #
d k
2!"n ! m , " 2n ! m
x! " 2!" nx# " "2n x $ 0 #
k
!n ! m natural frequency (undamped oscillator when d ! 0!
d
"! 2m! n
relative damping ratio
Damped Oscillator

For a damped system d > 0, the frequency of the oscilla>on will be smaller than the natural
frequency of the undamped system:
a
a 2 ! ! 2 " ! 2n !! "n ! cos# Im
a ! absolute damping factor jw

wn
! ! frequency of oscillation (damped system)
Eigenvalues -a=-n Re
! 1,2 ! ! "# n " j#
!a -jw
Undamped oscillator: a = 0
25


The upper plot shows a mass-damper-spring system for dierent rela>ve damping ra>os.
The lower plot shows the undamped oscillator together with a damped oscillator.
Matlab MSS
toolbox:
ExMDS.m
26


It is convenient to set:
! ! r !n
where r is a reduc>on factor.

For uncontrolled marine craK a reduc>on of 0.5 % percent in the natural frequency is quite
common (Fal>nsen 1990):
r ! 1! 0.5
! 0. 995 z = 0.1

a2 ! !r! n "2 " ! 2n #

100

d

m ! 2!" n
#
! 2 1 ! r2 "n
a" 1 ! r2 !n # k
" ! 2 1 ! r2 m #
which yields the following formula for linear damping:
d ! 2 1 ! r2 km Compute linear damping in heave,

roll, and pitch of an uncontrolled
marine craK using m and k.
27


Damping in surge, sway, and yaw can, however, not be determined by

d ! 2 1 ! r2 km
since k = 0 in a pure mass-damper system.

Linear damping for such a system:

mx! " dx# $ !

can be found by specifying the >me constant T > 0. Let T = m/d such that:

Tx! " x# $ 1d !

which yields the following design formula for the mass-damper:
m
d! T
28


Example 12.5 (Linear Damping in Roll and Pitch for Submarines)

Consider the linear pitch equa>on:

!I y ! Mq! "!" ! M q!! # BG z W ! $ " 5
Using:
d ! 2 1 ! r2 km
yields:

!M q ! 2 1 ! r2 BG z W!I y ! M q" " # 0

For roll a similar expression is obtained:

!Kp ! 2 1 ! r2 BG z W!I x ! Kp" " # 0

where r = 0.995 for = 0.1.
29


Example 12.6 (Linear Damping in Yaw for Ships and Underwater Vehicles)
Consider the linear yaw equa>on:

!I z ! N r! "r! ! N rr " N ! ! # mx! " !d " K d "x# " K p x $ 0 #
Assume that the moment of iner>a I z !
N
r! is known.
d " Kd
2!" n ! m #
The linear damping coecient Nr can be es>mated by specifying Kp
the >me constant.

"n ! m #

If we assume that the closed-loop yawing mo>on has a natural 2!" n ! 2! 2#
Tn
period Tn = 150 s and rela>ve damping ra>o = 1.0, we can
d " Kd
compute an es>mate of the >me constant in yaw using: ! m
! 2 #
T
T!
150 s
! 23. 8 s # T ! Tn #
2! ! 1. 0 2!"
Note that: Kd/m 1/T
corresponds to increasing
! N r ! I z ! N r" # 1/T to 2/T in closed loop

T
30


12.2.2 Acceleration Feedback
Consider a mass-damper-spring system: Feedback

mx! " dx# " kx $ ! " w ! ! ! PID ! K m x"
where Km > 0 is the accelera>on feedback gain and ! PID represents a conven>onal PID
controller. This yields:
!m ! Km "x" ! dx# ! kx $ ! PID ! w
x! " d
x# " k
x$

1

! " 1
w
m"Km m"Km m"Km PID m"Km
From this expression it is no>ced that besides increasing the mass

from m to m+Km
Accelera>on feedback also reduces the gain in front of the disturbance w from
1/m to 1/(m+Km). Hence, the system is expected to be less sensi>ve to an
external disturbances w if accelera>on feedback is applied.
31


This design can be further improved by introducing a frequency-dependent virtual mass
(Sagatun, Fossen and Lindegaard, 2001), that is:

! ! ! PID ! h m !s"x" The total mass is m+Km at low frequencies
If hm(s) is chosen as a low-pass lter: (s = 0) while at high frequencies (s = ) the
total mass m+Km reduces to m.
Km

h m !s" ! 1"T ms
it is seen that:

m ! Km
x" ! dx# ! kx $ ! PID ! w

1!T ms
m total!s"

The total mass is
Km mT ms"!m"K m"
m total !s" ! m " 1"T ms ! T ms"1
32


The lter hm(s) can be chosen rather arbitrarily depending on the applica>on.
1
x! " g!s"dx# " g!s"kx $ g!s"! PID " g!s"w g!s" ! m"hm!s"
For instance, a
low-pass lter
will remove high-
frequency
accelera>on

feedback
components
while a notch
structure can be
used to remove
1st-order wave-
induced
disturbances.
33


12.2.3 PID Control with Acceleration
Feedback
Consider the mass-damper-spring system:

mx! " dx# " kx $ !
t
! ! k d x d ! K px" # K dx$ # K i " o x" !!"d! ! h m!s"x%

reference
PID
acceleration

feedforward
control
feedback

For simplicity, assume that hm(s) = Km and Ki = 0. This gives the PD controller:

! ! k d x d ! !K px" # K dx$" ! K mx%

The closed-loop system becomes:
k " Kp
!n ! #
m " Km
!m ! Km "x" ! !d ! Kd"x# ! !k ! K p"x$ % w
" ! d " Kd #
2!m " K m "!n
34


Feedback
Pole placement (PD controller):
Let:
x! " 2!"n x# " "2n x $ 0

Solving for Kp and Kd using:

!m ! Km "x" ! !d ! Kd"x# ! !k ! K p"x$ % w
yields

Kp ! !m " Km"! 2n ! k, Kd ! 2"!n !m " Km" ! d

The disturbance w can be removed by adding integral ac>on.
35


Feedback
Pole placement (PID controller):
Consider: 1
! ! k d x d ! K p 1 " T d s " T is
x# ! Km x$

where Td = Kd/Kp and Ti = Kp/Ki are the deriva>ve and integral >me constants, respec>vely.
A rule-of-thumb is to choose:
1 !n
!
Ti

10

which states that the integrator is 10 >mes slower than the natural frequency ! n . This
yields:
!n !n 2
Ki ! 10 Kp ! 10 !"m " K m#! n ! k$
36


Feedback
The natural frequency !
n can be related to the system bandwidth !
b by using the
following deni>on:

Deni+on 12.1 (Control Bandwidth)
The control bandwidth of a system y=h(s)u with nega>ve unity feedback is dened as the
frequency !
b rad/s at which the loop transfer func>on l(s)=h(s)1 is:
2
|l!j!"|!!!b ! 2 u y
h (s )
or equivalently,

20 log|l!j!"| !!! b ! !3 dB

From this deni>on it can be shown that the control bandwidth of a 2nd-order system:
! 2n
h!s" ! s 2"2"! s"! 2 ! ! 1. 0
is n n

! b ! ! n 1 ! 2" 2 " 4" 4 ! 4" 2 " 2
! b ! 0. 64 ! n
37


Feedback
Main Result (PID Controller Tuning Rule):

t
! ! k d x d ! K px" # K dx$ # K i " x" !!"d! ! h m!s"x%
o
38


Feedback
Example 12.7 (Ship Autopilot Design)
Consider the rst-order Nomoto model:

T 1
T!! " !# $ K" m! K, d! K, k!0

Pole placement gives: Km = 0 (no angular accelera>on

feedback in yaw)
Km ! 0
!2n T
T " KK m 2 Kp ! "0
Kp ! !n # 0 K
K
2"! n T ! 1
K d ! T " KK m 2"!n " 1 # 0 Kd !
K
"0
K K
! 3T
K i ! T " KK m !3n # 0 Ki ! n " 0
10K
10K
39


12.2.4 MIMO Nonlinear PID Control with
Acceleration Feedback
The PID control concept can be generalized to nonlinear mechanical system by exploi>ng
the kinema>c equa>ons of mo>on in the design.

!" ! J!!"# #
M#" # C!#"# # D!#"# # g!!" ! $ % w #

Consider the control law:

! " g!#" ! Hm !s"$% ! J ! !#"! PID

with accelera>on feedback Hm(s), gravity compensa>on g, and PID controller:
t
! PID ! K p"# $ K d"% " K i ! o "#!!"d!

For simplicity, assume that Ki = 0 and Hm(s) = Km (PD control with xed gain accelera>on
feedback). This yields the closed-loop system:
H!! " !C"!# " D"!# " K !d ""#$! " J! !""K p "# # w K !d !!" ! J ! !!"K d J!!"
!" # ! ! !d H ! M ! Km
40


A Lyapunov func>on candidate (LFC) for this system is
1 ! 1 $! $ H ! H! ! 0
V! 2 ! H! " 2 # K p #
kine>c poten>al Kp ! K!p " 0
energy energy
In the analysis it is assumed that !"d ! 0, that is, regulation of ! to ! d ! constant.

V! " ! ! H!" # $" ! K p$% " ! ! !H!" # J ! !$"K p$%"
since #"! $ #! ! #!d $ #! and #! !$ % !J !!#".
Using:
H!! " !C"!# " D"!# " K !d ""#$! " J! !""K p "# # w ! ! C!!"! "0
V! " !! !w ! "C!!# # D!!# # K "d !"#$!! " ! !w ! ! !"D!!# # K "d !"#$!
41


If w = 0, Krasovskii-LaSalle's Theorem in Appendix A can be used to prove that the system with
nonlinear PD control (Ki = 0) is globally asympto>cally stable (GAS). Moreover, the trajectories
will converge to the set found from: !

V! !x! ! !" ! "D""! # K !d !$#$" " 0

which is true for ! = 0. Therefore:

! " !"!", #! : # $ 0#$
Now, ! 0 implies that H ! ! " !J

! !""K

p "# which is nonzero as long as !" 0.
Hence, the system cannot get stuck at an equilibrium point value other than ! " = 0
Since the equilibrium point (! " , #!
! "0, is the largest invariant set M in !
0! , the equilibrium
point is GAS if J(h) is dened for all h (no representa>on singulari>es).

In the case of a constant disturbance w 0, the system trajectories will converge to a ball
about the origin (0,0). The radius of the ball depends on the magnitude of the disturbance w.
This is referred to as uniform ul>mate boundedness (UUB).
42


12.2.5 Case Study: Heading Autopilot for
Ships and Underwater Vehicles
Block diagram showing a modern autopilot system
wind
loads
wind
feedforward
pilot

yaw angle
input and rate
reference control control compass
model system allocation and gyro
observer and
wave filter
43


The principal blocks of a heading angle autopilot system are:
PID Feedback Control: The control system provides the necessary feedback signal to track the
desired yaw angle d. The output is the yaw moment N.
Compass/Yaw Gyro: The compass measures the yaw angle which is needed for feedback. In
some cases a yaw gyro is available for yaw rate feedbackthat is, feedback from r .

Observer/Wave Filter: In its simplest form the 1st-order wave-induced mo>on components w
and rw are ltered out from the measurements y1= +w and y2 = r+rw, and consequently
prevented from entering the feedback loop. This is known as wave ltering where the output of

the lter is the LF mo>on components and r. This is necessary to avoid excessive rudder
ac>on. In cases where y2 is not measured the wave lter must be constructed as a state
observer so that r can be es>mated from the yaw angle measurement y1.

Wind Feedforward: In cases where a wind sensor is available for wind speed and direc>on, a
wind model can be used for wind feedforward. This is oKen advantageous since the integral
ac>on term in the PID controller does not have to integrate up the wind disturbance term.
However, an accurate model of the wind force and moment as a func>on of ship speed and
wind direc>on is needed to implement wind feedforward.
Reference Feedforward using a reference model for course-changing maneuvers. Course
keeping is obtained by using a constant reference signal.
Control Alloca>on: This module distributes the output from the feedback control system,
usually the yaw moment N, to the actuators (rudders and in some cases propellers and
thrusters) in an op>mal manner.
44


Course keeping Reference model for course changing (turning)
! d ! constant !! d " sat!r d " #

r! d " sat!a d " #
The main mo>va>on for using a rate a! d " !!2" # 1"# n sat!a d " ! !2" # 1"# 2n sat!rd " # #3n !! r ! ! d " #
limi>ng element in the reference
model is that the course-changing
maneuver will be described by
3 phases (posi>ve turn):
I Start of turn, .
accelera+on

(rd > 0 and 0 < rd amax )
II Steady turning
(rd = rmax and rd = 0)
III End of turn, decelera+on
.

(rd > 0 and amax rd < 0)
The same 3 phases applies to
nega>ve turns but with opposite
signs on rd and ad
45


Yaw Dynamics (Nomoto Model)
!I z ! N r! "r! ! N rr " ! wind # ! N #
r! " 1 r # m
T
1 !!
wind " ! N " # T ! m ! I z ! N r" # ! N ! !N " " #
d !N r

Control System (PID Controller with Reference and Wind Feedforward)
!! " ! ! ! d is the heading error
! N !s" ! !!" wind # ! FF !s"!K p 1 # Tds # 1 " $ !s" #
Tis
!PID !P ID
!n ! 1 !b #
1! 2" 2 " 4" 4 ! 4" 2 "2
Time domain:
Kp t K p ! m! 2n
! N !t" ! !!" wind # ! FF $
! K p " ! K p T d r$ !
T " 0 "$ !!"d! #
K d ! m 2"! n ! 1
i T"0
Kd # Ki # 2"! n m
T
!
Ki ! n Kp
10
46


Nomoto Model:
! 1 " m
1 ! PID #
T
r! " 1 r # m
1 !!
wind " ! N " #
T
1/m
h!s" ! ! e !s" ! #
Reference feedforward: PID s!1 " Ts"
Loop transfer func>on:
! FF ! m r" d # 1 r d #
T K p !T i T d s 2"T d s"1"
l!s" ! h!s"hPID !s" !
PID controller

Ti s 2!1"Ts"
1 T i T d s 2"T d s"1
hPID !s" ! Kp 1 " T ds " Ti s ! Kp Ti s
h PID !s" ! Kp 1 " Td s " 1

Ti s
1 " !Ti " Td "s " T d T i s 2
! Kp Ti ! Td such that T i " T i ! T d
Ti s
!1 " Ti s"!1 " Td s"
! Kp #
Ti s
47


Cx

Wind Feedforward
Vrw is the rela>ve wind speed
V rw ! u 2rw " v 2rw #
rw is the rela>ve wind angle of akack
! rw ! !atan2!v rw , u rw " # gr

urw and vrw are Cy
the

rela>ve wind veloci>es
u rw ! u ! u w ! u ! V w cos!! w ! "" #

v rw ! v ! v w ! v ! V w sin!! w ! "" # Anemometer measurements:

Wind speed and direc>on Vw and w
! wind " 1 " a V 2rw C N !# rw "ALw L oa #

2
Wind coecient CN in yaw
gr

Wind feedforward is frequently used since the integral Cn

ac>on term in the PID controller does not have to
integrate up the wind disturbance term.

However, an accurate model of the wind moment as a
func>on of measured wind speed and direc>on is gr

needed to successfully implement wind feedforward.
48


Control Alloca+on
The yaw moment can be generated by a single rudder:

! N ! !N " " #

or several actuators ui (i=1...r) sa>sfying

! N ! b! u,

u !!u 1 ,...,u r" ! #

For a rudder controlled craK, the input command is computed

! ! 1 "N #
!N !

In the case of several actuators we can use the generalized inverse to compute u if
the scalar bTb 0 (see Sec>on 12.3). This gives:
u ! b!b! b" !1 ! N #
49


12.2.6 Case Study: Heading Autopilot
with Acceleration Feedback for Ships and
Underwater Vehicles
The autopilot system can be extended to include accelera>on feedback by dieren>a>ng the
output of a yaw-rate gyro rgyro according to:
s
r! ! !f"s rgyro Limited dieren>ator
Yaw dynamics:
r! " 1 r # m
T
1 !!
wind " ! N " #

Control law:
t
$ ! K d r$ ! K i " "
! N ! !!" wind # ! FF ! K p " $ !!"d! ! K m r% # accelera>on feedback
0
! FF ! !m # K m " r% d # 1 r d #
T
Error dynamics (gains can be chosen using pole placement):

t No>ce the increased
!T ! Km " ! !1 ! Kd " ! K p e ! K i ! 0 e!!"d! " 0
moment of iner>a
50


12.2.7 Case Study: Cross-Tracking System
for Ships and Underwater Vehicles
Consider a path dened in terms of waypoints using the Cartesian coordinates (xk,yk) .

Waypoint guidance systems can be designed as trajectory tracking controllers. In its simplest
form this involves the use of a classical autopilot system where the yaw angle command d is
generated such that the cross-track error e is minimized.

s!t" ! !x!t" ! x k " cos!! k " " !y!t" ! y k " sin!! k " #

e!t" ! !!x!t" ! x k " sin!! k " " !y!t" ! y k " cos!! k " #

Path-xed reference frame {p} = (xp,yp,zp) with
origin in pk and x-axis rotated a posi>ve angle:

! k :! atan2!y k"1 ! yk ,x k"1 ! xk " #

Consequently, the error term e represents the
devia>on to the path in the y-direc>on
expressed in {n}

e ! y p #

51


12.2.7 Case Study: Cross-Tracking System
for Ships and Underwater Vehicles
A waypoint trajectory-tracking system is usually designed such that the craK can move
forward with reference speed Ud at the same >me as the path cross-track error is minimized.

The desired path can be generated using a route management system or by specifying the
desired route by waypoints. If weather data are available, the op>mal route can be
generated such that the eects of wind and water resistance are minimized.
v"0 and !"0

x! p " u cos!!" ! v sin!!" " U #
y! p " u sin!!" # v cos!!"

v"0 and !"0
" U!

#

e!s" ! h ! !s"!!s" " h b !s"b!s" #
h ! !s" ! e !s" ! 2 KU #
! s !1 " Ts"
y! p " U! #
h b !s" ! e !s" ! 2 U #
!! " r # b s !1 " Ts"
Tr! # r " K" # b # t
b! " 0 # ! ! !K p e ! K d ! K i " e!""d" #
0
e ! yp # Why do we need integral ac>on when using a

Nomoto model that already has a pure integrator?

52


12.2.8 Case Study: LOS Path-Following
Control for Ships and Underwater Vehicles
A line-of-sight (LOS) path-following controller can be designed for conven>onal craK by
adding a kinema>c control loop as shown in the gure.
wind
wind loads
feedforward
way-
points

North-East
positions
LOS control control
algorithm system allocation
Yaw rate
and angle
observer and
wave filter
autopilot
53


A marine craK can move along the LOS vector with dierent direc>ons of the BODY axes:

Body-Axes Alignment: If the sideslip angle is unknown, the body x-axis

of the craft can be aligned with the LOS vector to the price of a tracking offset
(see Section 10.3.2):

Enclosure-based Steering (Direct Assignment of the Course Angle d )

! d ! " d ! atan2!y los ! y, x los ! x" #

!d ! " d ! # #

Autopilot control system:
t
! N ! !!" wind # ! FF ! K p "$ ! K d "$% ! K i " "$ !!"d! #
0
! FF ! m r% d # 1 r d #
T
54


Body x-axis and LOS vector aligned

LOS guidance principle where the sideslip angle is chosen as zero
The price to be paid is that

the craK will behave like an
object hanging in a rope and
the craK's lateral distance to
the path will depend on the
magnitude of the
environmental forces and
thus the sideslip angle .
This is due to the fact that
= only if =0.
55


For lookahead-based steering, the course angle assignment is separated into two parts:
! d !e" ! ! p " ! r !e" #
Path-tangen>al angle w.r.t. NED: Time-varying lookahead distance: !"t# ! R 2 ! e"t# 2 #

!p ! "k #
e!t" 2 ! #!t" 2 " R 2 #
Velocity-path rela>ve angle:
!e!t"
! r !e" :! arctan
#
#

where > 0 is the lookahead
(carrot) distance.
56


Lookahead-Based Steering with Integral Ac+on:

! d ! ! p " ! r !e" #
Path-tangen>al angle:
!p ! "k #

Velocity-path rela>ve angle (P-type guidance law):

! r !e" :! arctan
!e!t"
#

#

Heading angle:
!d ! "d ! #
PI-type guidance law.
! "p " "r ! # # Integral ac>on compensates

If the sideslip angle is unknown, a PI-type guidance law: for unknown sideslip

t
! r ! arctan !K p e!t" ! K i " e!""d" # K p ! "1
0

together with the approxima>on d p+r can be used to compensate for sideslip .
57


12.2.10 Case Study: Dynamic Positioning of
Ships and Floating Structures
Control systems for sta>onkeeping and low-speed maneuvering are commonly known as
dynamic posi>oning (DP) systems. The are designed for simultaneously control of the three
horizontal mo>ons (surge, sway, and yaw) and they are used in a wide range of marine
opera>ons such as sta>onkeeping, drilling and ooading.

DNV (1990) denes a dynamically posi>oned vessel as a free-oa>ng vessel which maintains
its posi>on (xed loca>on or predetermined track) exclusively by means of thruster.

It is, however, possible to exploit rudder forces in
DP by using the propeller to generate rudder liK
forces.

The rst DP systems were designed using
conven>onal PID controllers in cascade with LP
and/or notch lters to suppress the wave-induced
forces. This was done by neglec>ng coupling terms
(Sargent and Cowgill 1976, Morgan 1978). From
the middle of the 1970's a new model-based
control concept u>lizing stochas>c op>mal control
theory and Kalman ltering techniques was
employed with the DP problem by Balchen (1976).
58


DP Model (VP coordinates)
! ! !!"wind " R ! !!"!PID #
!" p ! # !" ! R!#"$ # ! ! R!!"! p #
t
M#" " D# ! bp " $ " $ wind " $ waves #b ! R!!" ! b !PID ! K p "# " K d "$ " K i ! "#!!"d! #
p o
t
! ! !!"wind ! R ! !!"K p #$ ! R ! !!"K d R!!" % ! R ! !!"K i # #$!""d" # K !d :! R ! !!"K d R!!" #
K"d 0

For the full state feedback case, asympto>c stability
follows using Lyapunov arguments. However, in
order to implement the nonlinear PID controller a
state es>mator and wave lter must be designed;
see Chapter 11.

GAS and convergence of the nonlinear PID
controller in combina>on with a Kalman lter
cannot be guaranteed but the solu>on has been
used in many industrial systems with excellent
performance and robustness. Hence, from a
59
prac>cal point of view this is a well proven concept.

In commercial DP systems it is necessary to include the following features:

v Integral ac+on to compensate for slowly varying disturbances (bias
term b) due to ocean currents and wave driK forces (2nd-order wave-
induced forces).

v Wind feedforward control to compensate for mean wind
disturbances. Wind gust cannot be compensated for since the
actuators do not the have the capacity of moving a large vessel in the
frequency range of the wind gust.

v Wave ltering to avoid that 1st-order wave-induced mo>ons fed back
to the control system. This is an important feature since the actuators
cannot move a large vessel fast enough to suppress the disturbances.

v State es+mator for noise ltering and es>ma>on of unmeasured
states, for instance linear and angular velocity. The main tool for this is
the Kalman lter, alterna>vely nonlinear and passive observers.

v Op+mal control alloca+on of thrust where the main goal is to
compute op>mal set-points for thrusters, rudders and other actuators
based on the force/moment commands generated by the DP control Copyright Bjarne Stenberg
60
system.


61


Wind Feedforward
CX

t
! ! !!"wind ! R ! !!"K p #$ ! R ! !!"K d R!!" % ! R ! !!"K i # #$!""d" #
K"d 0
C X !" rw "AF w grw

!" wind ! 1 ! a V 2rw C Y !" rw "AL w #
2 CY

C N !" rw "AL w L

V rw ! u 2rw " v 2rw #
! rw ! !atan2!v rw , u rw " # grw

u rw ! u ! u w ! u ! V w cos!! w ! "" # CN

v rw ! v ! v w ! v ! V w sin!! w ! "" #
grw

62

!" p ! # #
! ! !K d " #
M#" " D# " K p ! p ! bp " $ " $wind " $ waves #
The addi>onal spring Kp due to the mooring system adds spring s>ness in surge, sway and yaw
described by the parameters k > 0, k > 0 and k 0.
M ! M! !
m 11
0
0
m 22 m 23
0
#

0 m 32 m 33
d 11 0 0
D ! D! ! 0 d 22 d 23 #
0 d 32 d 33
K ! diag!k 11 , k 22 , k 33 " #
63

12.2.11 Case Study: Position Mooring System
for Ships and Floating Structures
Turret mooring systems have cables that are connected to the turret via bearings. This allows
the vessel to rotate around the anchor legs and the rota>onal spring can be neglected (k = 0).
The turret can be mounted either internally or externally. An external turret is xed, with
appropriate reinforcements, to bow or stern of the ship. In the internal case the turret is placed
within the hull in a moon pool. Turret mooring systems allows the vessel to rotate in the
horizontal plane (yaw) into the direc>on where environmental loading due to wind, waves and
currents is minimal. This is referred to as weathervaning.

Spread mooring systems are used to moor

Floa>ng Produc>on, Storage and Ooading

(FPSO) units, tankers and oa>ng plaorms. The
system consists of mooring lines akached
somewhere to the vessel. The drawback with a
spread mooring system is that it restrains the
vessel from rota>ng (k > 0) and hence
weathervaning is impossible.
64

12.2.11 Case Study: Position Mooring
System for Ships and Floating Structures
For thruster assisted PM systems the thrusters are
complementary to the mooring system and the main
idea is to provide the system with addi>onal damping
by using a D-controller:

! ! !K d " #

Integral ac>on is not used in PM systems, since the ship

is allowed to move within a limited radius about the

equilibrium point or eld-zero point (FZP).
If the vessel, moves outside the specied radius of the
mooring system, a stabilizing control system of PD-type
can be used to drive the vessel inside the circle again. It
is op>mal to use addi>onal thrust to stay on the circle
rather than move the vessel to the FZP. In good
weather, no control ac>on is needed since the vessel is
free to move within the circle.

65


12.3 Control Allocation
For marine craK moving in n DOF it is necessary to distribute the generalized control forces
! ! ! n given by:

! " Bu

to the actuators in terms of control inputs u ! ! r
From Section 9.4:

r n fully actuated control
r < n underactuated control

The input matrix B is square for r = n, that is the number of actuators is equal to n DOF.
Computa>on of u from ! is a model-based op>miza>on problem which in its simplest form

is unconstrained.
Physical limita>ons like input amplitude and rate satura>ons imply that a constrained
op>miza>on problem must be solved.

66

Lecture Notes TTK 4190 Guidance and Control (T. I. Fossen)

12.3.1 Actuator Models
Some actuators that can be rotated at the same >me as they produce control forces.
! " B!#"u !"!! 1 , . . . , ! p! ! , u""u 1, . . . , ur ! !

vector of angles
Rotatable Thrusters and Propellers
Examples are azimuth thrusters on an oshore supply vessel, podded propellers that can be
rotated, and contra-rota>ng propellers.
This increases the number of available controls from r to r+p, where p denotes the number
of rotatable actuators for which addi>onal nonlineari>es are introduced.

67


Main propellers: the main propellers of the vessel are mounted aK of the hull usually in
conjunc>on with rudders. They produce the necessary force Fxi in the x-direc>on
needed for transit.

Tunnel thrusters: transverse thrusters going through the hull of the vessel. The propeller
unit is mounted inside a transverse tube and it produces a force Fyi in the y-direc>on.
Tunnel thrusters are only eec>ve at low speed which limits their use to low-speed
maneuvering and dynamic posi>oning.

tunnel thrusters
Fx5
main
Fy3 Fy2 Fy1 x
propellers
Fx4
68


Azimuth thrusters: thruster units that can be rotated an angle
about the z-axis and produce two force components (Fx1, Fy1)
in the horizontal plane are usually referred to as azimuth
thrusters.
Azimuth thrusters are usually mounted under the hull of the
vessel and the most sophis>cated units are retractable.
Azimuth thrusters are frequently used in DP since they can
produce forces in dierent direc>ons leading to an
overactuated control problem that can be op>mized with

respect to power and possible failure situa>ons.

Fx1

x

Fy1

F1

y

69


Stabilizing ns: stabilizing ns are used for damping of
ver>cal vibra>ons and roll mo>ons. They produce a force Fzi
in the z-direc>on which is a func>on of the n deec>on.
For small angles this rela>onship is linear. Fin stabilizers can
be retractable allowing for selec>ve use in bad weather. The
liK forces are small at low speed so the most eec>ve
opera>ng condi>on is in transit.

Ad rudders: rudders are the primary steering device for conven>onal vessels.
They are located aK of the vessel and the rudder force Fy1 will be a func>on of
the rudder deec>on. A rudder force in the y-direc>on (liK is perpendicular to
speed U while drag is in the direc>on of the speed) will produce a yaw
moment which can be used for steering control.
Fy1
x

Rudder force
1
Fy1 = CL ( ) ARU 2 sin( R ) k R y

70
2

Control surfaces: control surfaces can be mounted at dierent loca>ons to
produce liK and drag forces. For underwater vehicles these could be ns for
diving, rolling, and pitching, rudders for steering, etc.
Water jets: water jets is an alterna>ve to main propellers aK of the ship. They
are usually used for high-speed craK.
71


The control force due to a propeller, a rudder, or a n is (assuming linearity):
k is the force coecient

F! ku u is the control input depending on the actuator considered
Forces and moments in 6 DOF corresponding to the force vector

f ! !Fx , Fy, Fz! !

x , l y , l z ! ! can be wriken:
and moment arms r ! !l

Fx
Fy Fx
f Fz 4 DOF Fy
!! ! ! !!
r!f F z l y ! Fy l z Fz l y ! Fy l z
F x l z ! Fz l x Fy l x ! Fx l y
F y l x ! Fx l y
72


Example (Linear model describing nonlinear monotonic control forces):
Consider the linear model:
F! ku
This model can also be used to describe nonlinear monotonic control forces.
For instance, if the rudder force F is quadra>c in rudder angle , that is:

F ! k !|!|
Choosing
u ! !|!| ! ! sign!u" |u|
and the resul>ng model is linear in u.
73


The linear control force F = ku for the dierent actuators can be represented in terms of a
force vector f according to the following table:
Actuator u (control input) ! (control input) f ! (force vector)

main propellers (longitudinal) pitch and rpm !F, 0, 0"
tunnel thrusters (transverse)

pitch and rpm

!0, F, 0"
azimuth (rotatable) thruster pitch and rpm angle !F cos !, F sin !, 0"
aft rudders angle !0, F, 0"
stabilizing fins angle !0, 0, F"
74


The control forces and moments f = [f,...,fn]T are conveniently expressed as:
f ! Ku #
Force coecient matrix

The force coecient matrix K is diagonal such that:

K ! diag!K 1 , . . . , K r", K !1 ! diag 1 ,..., 1 #

K1

Kr
The actuator forces and moments relate to the control forces and moments by:

! ! T!""f T() is the thrust congura>on matrix.
! T!""Ku #

where u and are control inputs dened by the vectors:
!"!! 1 , . . . , ! p! ! , u""u 1, . . . , ur ! !
75


Thrust congura+on matrix for nonrotatable thrusters:
The thrust congura>on matrix is dened in terms of a set of column vectors t . n
i !!
T ! !t 1 , . . . , t r" ! constant
In 4 DOF (surge, sway, roll, and yaw) the column vectors take the following form:
1 0

0
0 1 0
ti ! , ti ! , ti !
0 !l zi l yi
l yi l xi 0
tunnel thruster
main propeller stabilizing fin
and aft rudder
76


Example: 3-DOF mo+ons (surge, sway and yaw):
0 0 0 1 1
T! 1 1 1 0 0 # K ! diag!K 1 , K 2 , K 3 , K 4 , K 5 " #
l x1 l x2 l x3 l y4 l y5

l
x1
l x3
l x2
Fx5
l y5 Fy3 Fy2 Fy1 x

l y4
Fx4
main y
tunnel thrusters
propellers
77


Thrust congura+on matrix for rotatable thrusters:
For marine craK equipped with azimuth thrusters in combina>on with nonrotatable
thruster:
T!!!" "t 1, . . . , t r"
In 4 DOF (surge, sway, roll, and yaw) the column vectors take the following form:

cos !! i " 1 0 0
sin !! i " 0 1 0
ti ! , ti ! , ti ! , ti !
!l zi sin !! i " 0 ! l zi l yi
l xi sin !! i " ! l yi cos !! i " l yi l xi 0
azimuth main tunnel thruster stabilizing
thruster propeller and aft rudder fin
78


Extended thrust congura+on matrix for rotatable actuators:
The rotatable thrusters can be treated as two forces. Consider a rotatable thruster in the
horizontal plane (the same methodology can be used for thrusters that can be rotated
in the ver>cal plane):
F xi ! F i cos!! i " F yi ! F i sin!! i " Fx1

! K i u i cos!! i " # ! K i u i sin!! i " #
Extended force vector: Fy1

f e ! K e ue #

F1

! ! Te K e ue #

where Te and Ke are the extended thrust congura>on and coecient matrices,
respec>vely and ue is a vector of extended control inputs where the azimuth controls are:
u ix ! u i cos!! i " #
u iy ! u i sin!! i " #
79


Example 12.8 (Thrust congura+on matrix for an underwater vehicle):
The forces and moment in surge, sway and yaw, satisfy:
! ! T!""Ku #
"
K1 0 0 0 u1
X cos!! 1 " cos!! 2 " 1 1
0 K2 0 0 u2
Y ! sin!! 1 " sin!! 2 " 0 0 #
0 0 K3 0 u3
N l x1 sin!! 1 " l x2 sin!! 2 " !l y3 !l y4
0 0 0 K4 u4

NB! This expression is nonlinear in i

80


Example 9.14 (Cont) Extended thrust configuration representation:
! e ! Te K e u e #
"
K1 0 0 0 0 0 u 1x
0 K1 0 0 0 0 u 1y
X 1 0 1 0 1 1
0 0 K2 0 0 0 u 2x
Y ! 0 1 0 1 0 0 #
0 0 0 K2 0 0 u 2y
N 0 l x1 0 l x2 !l y3 !l y4
0 0 0 0 K3 0 u3
thrust ue can be found by matrix inversion

NB! This is a linear problem where the extended 0

0 0 0 0 K4 u4
u1 ! u 21x " u 21y , ! 1 ! atan2!u 1y , u 1x "

81

u2 ! u 22x "
u 22y , ! 2 ! atan2!u 2y , u 2x "
12.3.2 Unconstrained Control Allocation
for Nonrotatable Actuators
The simplest alloca>on problem is the one where all control forces are produced by
thrusters in xed direc>ons alone or in combina>on with rudders and control surfaces
such that:
! " T!#"Ku ! " ! 0 ! constant, T " T!! 0 "
We will assume that the alloca>on problem is unconstrained

For marine craK where the congura>on matrix T is square (r = n) or nonsquare (r > n),
that is there are equal or more control inputs than controllable DOF, it is possible to nd
an op>mal distribu>on of control forces:
f = Ku

for each DOF by using an explicit method.
Solu+on: Unconstrained least-squares (LS) op>miza>on problem
82


Unconstrained least-squares (LS) op+miza+on problem (Fossen 1991)
J ! min !f ! Wf" minimizing force f = Ku

f
subject to: ! ! Tf ! 0
W is a posi>ve denite matrix, usually diagonal, weigh>ng the control forces f.

Lagrangian
L!f, !" ! f ! Wf " ! ! !" ! Tf" ! ! ! r is a vector of Lagrange multipliers
!L ! 2Wf " T! ! ! 0 # ! ! Tf ! 1 TW!1 T! " #

!f 2
" "
f ! 1 W"1 T! ! # " ! 2!TW!1 T! " !1 ! # assuming that
2 !1
TW T
! is nonsingular
Lagrange mul>pliers
83


Lagrange mul>pliers:
! ! 2!TW!1 T! ! !1 "
Op>mal force vector:
f! 1
2 W!1 T! !

Generalized inverse:
f ! W!1 T! !TW!1 T! ! !1 ! #
" T!w ! # T!w ! W!1 T! !TW!1 T! ! !1
For the case W = I, that is equally weighted
Control alloca>on control forces, reduces to the

Moore-Penrose pseudoinverse:
f = Ku u ! K !1 T!w "
T! ! T" !TT" "!1
84


12.3.3 Constrained Control Allocation for
Nonrotatable Actuators
In industrial systems it is important to minimize the power consump>on by taking advantage of
the addi>onal control forces in an overactuated control problem.

From a safety-cri>cal point of view it is also important to take into account actuator limita>ons
like satura>on, wear and tear as well as other constraints.

In general, this leads to a constrained op>miza>on problem.

This problem can be solved explicit (no itera>ons) using Piece-Wise Linear Func>ons

P. Tndel, T. A. Johansen and A. Bemporad. An algorithm for mul>-parametric quadra>c
programming and explicit MPC solu>ons, IEEE Conf. Decision and Control, Orlando, Vol. 2,
pp. 1199-1204, 2001
An on-line algorithm is presented in:
P. Tndel, T. A. Johansen and A. Bemporad. An algorithm for mul>-parametric quadra>c
programming and explicit MPC solu>ons, Automa)ca, Vol. 39, pp. 489-497, 2003
85


J !min !f ! Wf ! s ! Qs !!f"" Op+miza+on criterion
f,s,f"
subject to:
Tf ! " # s s is a vector of slack variables

that should be close to zero, that is:
f min ! f ! f max
Q ! W !0
"f" ! f 1 , f 2 , . . . fr ! "f

The 1st term of the criterion corresponds to the LS criterion (pseudo-inverse)
The 3rd term is introduced to minimize the largest force !f " max i |f i | among the
actuators.
The constant ! ! 0 controls the rela>ve weigh>ng of the two criteria.
min max
This formula>on ensures that the constraints f i ! f i ! f i (i ! 1,...,r!
are sa>sed, if necessary by allowing the resul>ng generalized force Tf to
deviate from its specica>on.

86


Global Solu+on:
The global solu>on can be computed o-line using mul>-parametric QP algorithms;
see Tndel et al. (2001, 2003).
Local Solu+ons:
Explicit Solu+ons based on Minimum Norm and Null-Space Methods: In ight and
aerospace control systems, the problems of control alloca>on and satura>ng control

have been addressed by Durham (1993, 1994a, 1994b).
They also propose an explicit solu>on to avoid satura>on referred to as the direct method.
By no>cing that there are innite combina>ons of admissible controls that generates
control forces on the boundary of the closed subset of akainable controls, the direct
method calculates admissible controls in the interior of the akainable forces as scaled
down versions of the unique solu>ons for force demands.
87


Case Study: CyberShip I

Explicit optimal constrained least-squares solution (nonrotatable thrusters)

Supply vessel scale 1:70 equipped with
4 azimuth thruster. The controls are:

RPM controlled
Azimuth angles
Azimuth angles are xed in the experiment

Problem: compute the op>mal controls

u = [u1,u2,u3,u4]T when u3 saturates at 0.1 N.

NB! Cannot be solved using the Generalized Inverse
Reference:
T. A. Johansen, T. I. Fossen, P. Tndel, Ecient Op>mal Constrained Control Alloca>on
via Mul>-Parametric Programming. AIAA Journal of Guidance, Dynamics and Control,
2005.
88


Least-squares thrust allocation (generalized inverse)

In the case of no satura>on the control alloca>on problem is solved as:
89


Constrained least-squares thrust allocation
An explicit piecewise linear representa>on of the solu>on can be found using

Mul>parametric Quadra>c Programming (mp-QP)
90


DP Experiments: Slowly Increasing Wind Loads
The increase in wind forces increases the demand for thrust and possible
satura>on in the control u
91


92


Azimuth thruster no. 3 hits satura>on limit when wind

force increases. Solu>on is s>ll OPTIMAL
93


DP Experiments: Position Change
94


DP experiments: Position Change
95


Azimuth Thrusters
Itera+ve Solu+ons: An alterna>ve to the explicit solu>on could be to use an itera>ve

solu>on to solve the QP problem.

The m-le func>on quadprog.m in the Matlab op>miza>on toolbox can be used for
computer simula>ons, while a stand-alone compiled QP solver must be implemented in
a real->me applica>on.
The drawback with the itera>ve solu>on is that several itera>ons may have to be
performed at each sample in order to nd the op>mal solu>on.
An advantage of the itera>ve approach is that there is more exibility for online
recongura>on, as for example a change in W may require that the explicit solu>ons
are recalculated. Computa>onal complexity is also greatly reduced by a warm start,
that is the numerical solver is ini>alized with the solu>on of the op>miza>on problem
computed at the previous sample.
96


Azimuth Thrusters
The control alloca>on problem for vessels equipped with azimuth thrusters is in general a
nonconvex op>miza>on problem that is hard to solve.
The primary constraint is:
! ! T!""f ! ! ! p denotes the azimuth angles
The azimuth angles a must be computed at each sample together with the
control inputs u

Other problems:
feasible sectors ! i,min ! ! i !! i,max

limiting turning rate !"
T!w !!" ! W!1T!!!"#T!!"W!1T!!!"# !1 can be singular for certain !-values
97


Azimuth Thrusters
r
!
J ! min
f,!,s
! P" i |f i |3/2 # s! Qs # !! " ! 0" !"!! " ! 0" # " # det!T!!"W "1 T! !!""
#
i!1
subject to: power avoid singular

consump>on congura>ons
T!!"f # $ # s
f min # f # f max
ensures that the azimuth
! min # ! ! ! max
$! min # ! " ! 0 ! %! max

angles do not move to
much within one sample
This op>miza>on problem is a nonconvex nonlinear program and it requires a signicant

amount of computa>ons at each sample.
Two alterna>ve implementa>on strategies will be men>oned.
T. A. Johansen, T. I. Fossen, and S. P. Berge. Constrained Nonlinear Control Alloca>on

With Singularity Avoidance Using Sequen>al Quadra>c Programming. IEEE Transac)ons on
Control Systems Technology, Vol. 12 , No. 1, Jan. 2004, pp. 211-216
98


Azimuth Thrusters
Itera+ve solu+ons using quadra+c programming
The nonlinear program can be locally approximated with a convex QP problem by assuming
that:
1. Power consump>on can be approximated by a quadra>c term in f,
near the last force f0 such that f = f0 + f
2. The singularity avoidance penalty can be approximated by a linear term linearized
about the last azimuth angle ! 0 such that ! ! ! 0 " #!

The resul>ng QP criterion is (Johansen, Fossen and Berge, 2004):
!
J ! min !f 0 # "f" ! P!f 0 # "f" # s ! Qs # "" ! #"" # ! "" #
"f,!",s !" " # det!T!""W"1 T! !""" "0
subject to:
!
s $ T!" 0 "!f $ !" !T!" 0 "f " " 0,f 0
!" ! % " T!" 0 "f 0
f min " f 0 # f # f max " f 0
" min " " 0 # !" ! " max " " 0
"" min # !" ! !" max
99


Azimuth Thrusters
Itera+ve solu+ons using linear programming:
Linear approxima>ons to the thrust alloca>on problem has been discussed by
Webster (1999) and Lindfors (1993).
In Lindfors (1993) the azimuth thrust constraints:
|fi | ! !fi cos !i "2 " !fi sin ! i " 2 ! fmax

i

are represented as circles in the !f
i cos !
! i , f i sin i " -plane.
The nonlinear program is transformed to a linear programming (LP) problem by
approxima>ng the azimuth thrust constraints by straight lines forming a polygon. If 8 lines
are used to approximate the circles (octagons), the worst case errors will be less than 4.0%.
However, singulari>es and azimuth rate limita>ons are not weighted in the cost func>on
100


Azimuth Thrusters
Explicit solu+on using the singular value decomposi+on and ltering techniques:

An alterna>ve method to solve the constrained control alloca>on problem is to use the
singular value decomposi>on (SVD) and a ltering scheme to control the azimuth
direc>ons such that they are aligned with the direc>on where most force is required,
paying aken>on to singulari>es (Srdalen et al. 1997).

Results from sea trials have been presented in (Srdalen et al. 1997). A similar technique
using the damped-least squares algorithm has been reported in Berge and Fossen
(1997) where the results are documented by controlling a scale model of a supply
vessel equipped with four azimuth thrusters.
101


Case Study: Nonlinear Programming Approach:
Power Optimality and Singularity Avoidance
Constraints:
T. A. Johansen, T. I. Fossen, and S. P. Berge. Constrained Nonlinear Control Alloca>on With

Singularity Avoidance Using Sequen>al Quadra>c Programming. IEEE Transac)ons on Control
Systems Technology, Vol. 12 , No. 1, Jan. 2004, pp. 211-216
102


Case Study: Nonlinear Programming Approach:
Power Optimality and Singularity Avoidance
Local convex quadra>c approxima>on at each sample

standard QP solu>on
Tradeo between maneuverability and power op>mality
singularity avoidance
Dynamic solu>on that handles rate and amplitude constraints

We will consider a marine craK with four azimuth thrusters (-25o to 25o azimuth,
1 deg/s rate constraints)
Scenario: Constant forward thrust plus course-changing maneuver
1. Op>mal approach with singularity avoidance

2. Op>mal approach without singularity avoidance
3. Pseudo-inverse
103


Simulation Example
Blue op>mal

Green op>mal
without singularity
avoidance

Red pseudo-inverse
104


Simulation Example
Blue op>mal

Green op>mal
without singularity
avoidance

Red pseudo inverse
105


Simulation Example
Blue op>mal

Green op>mal
without singularity
avoidance

Red pseudo inverse
106


Simulation Example
Blue op>mal

Green op>mal
without singularity
avoidance

Red pseudo inverse
107


12.3.5 Case Study: DP Control Allocation
System
Most DP ships use thrusters to maintain their posi>on and heading.
Fixed-Pitch (FP) propellers:

The thrust F from a variable speed FP
propeller can be modelled as:

F!n" ! Kn|n|
F!n" ! Kn

where K = constant is the thrust
coecient and n is the propeller
revolu>ons per minute (rpm).

Some propellers show a linear
behavior in n while others are
quadra>c. Even combina>ons of the
linear and quadra>c behavior are
observed in prac>ce.
108


System
Most DP ships use thrusters to maintain their posi>on and heading.
Controllable-Pitch (CP) propellers:

CP propellers are screw blade propellers
where the blades can be turned under the
control of a hydraulic servo. This introduces
a second control variable, pitch p, which is
used to obtained the desired thrust F for

dierent propeller revolu>ons n.

P is the "traveled distance per revolu>on
D is the propeller diameter
p = P/D represents the pitch ra>o

The thrust from a xed speed CP propeller
can be approximated by:

F!n,p" ! K!n"|!p ! p 0 "|!p ! p 0 "

F!n,p" ! K!n"!p ! p 0 "

The pitch oset is denoted as p

109


System
Example 12.9 (experimental thrust characteris+cs)
We will consider a supply vessel equipped with main propellers and tunnel thrusters.

Func>ons found by curve gng:
main propeller
F!122, p" ! 370|p|p
F!236, p" ! 137|p|p

tunnel thruster F!160, p" ! 655|p|p
Measured thrust for a supply vessel equipped with main and tunnel thrusters. Asterisks denote
the measured data while the solid lines are least-squares ts to the quadra>c thrust func>on .
110


System
Example 12.10 (supply vessel thrust congura+on and force coecient matrices)
Computa>on of T() can be illustrated by considering a supply vessel equipped with two
main propellers (aK of the ship), two tunnel thrusters and two azimuth thrusters which can be
rotated to arbitrary angles and , and therefore produce thrust in dierent direc>ons

! ! T!""f # CP: u ! !|p 1 |p 1 , |p 2 |p 2 , . . . , |p r |p r " ! , (or u ! !p 1 , p 2 , . . . , p r " ! )
#
f ! Ku # FP: u ! !|n 1 |n 1 , |n 2 |n 2 , . . . , |n r |n r " !, (or u ! !n 1 , n 2 , . . . , n r " !)

We have 8 control variables (6 rpm set-points and 2 azimuth angles) for 3 DOF.
111


System

u 1 , ! 1 fore azimuth thruster u 4 aft tunnel thruster
u2 fore tunnel thruster u 5 starboard main propeller
u 3 , ! 2 aft azimuth thruster u 6 port main propeller
K ! diag!K 1 , K 2 , K 3 , K 4 , K 5 , K 6 " #
cos#! 1 $ 0 cos#! 2 $ 0 1 1
T#!$ ! sin#! 1 $ 1 sin#! 2 $ 1 0 0 #
l 1 sin#! 1 $ l 2 l 3 sin#! 2 $ l 4 !l 5 !l 6
112


System
DP Thrust congura>on matrices for supply vessel:
! ! T!""Ku #
"
X cos!! 1 " 0 cos!! 2 " 0 1 1
Y ! sin!! 1 " 1 sin!! 2 " 1 0 0
N l 1 sin!! 1 " l 2 l 3 sin!! 2 " l 4 !l 5 !l 6

K 0

0
0 0 0 u1
1
0 K2 0 0 0 0 u2
0 0 K3 0 0 0 u3
# #
0 0 0 K4 0 0 u4
0 0 0 0 K5 0 u5
0 0 0 0 0 K6 u6
113


System
Example 12.11 (supply vessel thrust alloca+on)
In order to implement a DP control system for the supply a thrust alloca>on algorithm is
needed. The simplest algorithm is the generalized inverse:

u ! K !1 T! !!"" #

T! !!" ! W!1 T" !!"#T!!"W!1 T" !!"$ !1 # most avoid singular combina>on of i

where W = WT > 0 is a posi>ve denite weigh>ng matrix, usually chosen to be diagonal.

W should be selected so that using the tunnel and azimuth thrusters is less expensive
(small Ki-value) than using the main propellers (large Ki-value). This solu>on is easy to
use for constant azimuth angles . As soon as is allowed to vary or the control inputs
saturates, an strategy for this must be developed. This signicantly complicates the
control alloca>on soKware.

For system with quadra>c thrust characteris>cs, the computed ui-values must be
mapped to pitch or rpm commands. If ui = |pi|pi, it is straighorward to verify that:
p i ! sign!u i " u i #
114


Control For Control Sustem

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Control For Control Sustem

Uploaded by

Copyright:

Available Formats

Chapter 12 Motion Control Systems

12.1 Open-Loop Stability and Maneuverability

Chapter 12 covers state-of-the-art PID

Copyright Bjarne Stenberg

x! " u cos ! ! v sin ! " u 0 cos ! # cruise speed u0

PD controller ! ! !Kp !" ! " d" ! Kd r

Resul>ng ! " !1 " KKd "!# " KKp ! $ KKp ! d " w #

! " d!# " k! $ f!t"

Matlab MSS toolbox script: Stabdemo.m

XY-Plot: Straight-line stability

Two types of direc>onal stability are

1. No Oscilla+ons: (d - 4km 0): This r (deg/s)

implies that both eigenvalues are 0.06 0.7

nega>ve and real-i.e. 1 such that:

!d " d ! 4km 0.02 0.3

! !" " "2 ! 1 #n # 0 -0.01

!d " j 4km ! dm XY-Plot: Directional stability (underdamped)

no restoring forces in yaw.

are present (k > 0), no feedback is 0 0

required to damp out the

A PID controller can be

The Norrbin Measure of Maneuverability

! ! !t! " " K ! #t ! ! T ! # T ! exp!!!t ! /T ! ""$" ! !t ! "

The ship's turning radius R is dened as:

This gives the non-dimensional formula for the turning radius:

R Y "v !N "r !m " x "g "!N "v !Y "r !m " " 1

If r ! rss is constant in steady-state, that is, r" ! r# ! !# ! 0 directly gives ! ss ! H B !r ss "

a ! absolute damping factor jw

a2 ! !r! n "2 " ! 2n #

which yields the following formula for linear damping:

d ! 2 1 ! r2 km Compute linear damping in heave,

Example 12.5 (Linear Damping in Roll and Pitch for Submarines)

!m ! Km "x" ! dx# ! kx $ ! PID ! w

From this expression it is no>ced that besides increasing the mass

In the analysis it is assumed that !"d ! 0, that is, regulation of ! to ! d ! constant.

since #"! $ #! ! #!d $ #! and #! !$ % !J !!#".

V! " !! !w ! "C!!# # D!!# # K "d !"#$!! " ! !w ! ! !"D!!# # K "d !"#$!

Block diagram showing a modern autopilot system

! d ! constant !! d " sat!r d " #

!I z ! N r! "r! ! N rr " ! wind # ! N #

h PID !s" ! Kp 1 " Td s " 1

! wind " 1 " a V 2rw C N !# rw "ALw L oa #

Wind feedforward is frequently used since the integral Cn

Error dynamics (gains can be chosen using pole placement):

v"0 and !"0

y! p " u sin!!" # v cos!!"

e ! yp # Why do we need integral ac>on when using a

The price to be paid is that

Path-tangen>al angle w.r.t. NED: Time-varying lookahead distance: !"t# ! R 2 ! e"t# 2 #

C X !" rw "AF w grw

Computa>on of u from ! is a model-based op>miza>on problem which in its simplest form

! " B!#"u !"!! 1 , . . . , ! p! ! , u""u 1, . . . , ur ! !

k is the force coecient

Forces and moments in 6 DOF corresponding to the force vector

and the resul>ng model is linear in u.

Actuator u (control input) ! (control input) f ! (force vector)

Force coecient matrix

F xi ! F i cos!! i " F yi ! F i sin!! i " Fx1

Extended force vector: Fy1

thrust ue can be found by matrix inversion

u1 ! u 21x " u 21y , ! 1 ! atan2!u 1y , u 1x "

! " T!#"Ku ! " ! 0 ! constant, T " T!! 0 "

We will assume that the alloca>on problem is unconstrained