Marine Vehicle Dynamics and Control

Mar i ne vessel mot i on i s mor e compl i cat ed

t han spacecr af t mot i on due t o t he dependence
of f or ces and moment s on at t i t ude and
vel oci t y
Mar i ne vessel mot i on i s i n some ways mor e
compl i cat ed t han ai r cr af t
mot i on
Typi cal f i r st i nvest i gat i ons
of shi p dynami cs and cont r ol
f ocuses on t he hor i zont al
mot i on of symmet r i c
( l ef t - t o- r i ght ) shi ps
( sur ge, sway, and yaw)
Reference Frames for Ship D&C
Inertial Frame, F
i
. Typically assume at Earth, but rotating
Earth is used for some problems. The inertial frame has 3-axis
down, though some sources have 3-axis up
For maneuvering problems, an Earth-xed F
i
is used, and for
seakeeping problems, a frame moving with the vehicles nominal
velocity is used
Body Frame, F
b
. A 3-2-1 rotation from F
i
through yaw (),
pitch (), and roll () angles: R
bi
= R
1
()R
2
()R
3
()
Reference Frames
for Ship D&C (2)
Illustration (from T. Fos-
sens Guidance and Control of
Ocean Vehicles) shows rela-
tionship between intermediate
axis systems and roll, pitch,
and yaw, angles
Motion Equations for a Rigid Ship
F
b
: v =

v +
1
m

f
uid
+f
thrust

+R
bi
a
grav
F
b
: = I
1

I +g
uid
+g
thrust

F
i
: r = R
ib
v

= S
1
()
where a
grav
= [0 0 g]
T
Note that v and are expressed in F
b
, and r is expressed
in F
i
Much of the ship d & c literature follows aircraft notation:
v = [u v w]
T
, = [p q r]
T
Assuming left-right symmetry (but not up-down symme-
try) implies that I
xy
= I
yz
= 0, but in general I
xz
6= 0
Motion Equations for a Rigid Ship
Equations of motion are typically written in this form
M +C() +D() +g() =
E
+
where = [v
T

T
]
T
, = [r
T

T
]
T
, M is the modied
mass matrix, C is a matrix involving the gyroscopic terms,
D is the damping matrix, g is the vector of restoring forces
and moments,
E
is the vector of environmental forces and
moments, and is the vector of control forces and moments
Determining the various matrices, forces and moments, is
rather complicated and is the subject of texts such as
T. I. Fossen, Guidance and Control of Ocean Vehicles, Wi-
ley, 1994
M. S. Triantafyllou and F. S. Hover, Maneuvering and
Control of Marine Vessels, Department of Ocean Engineer-
ing, MIT, 2003 (available online)
Modeling Fluid Forces
Assumption: Forces and moments acting on a rigid body are a
linear combination of
1. Radiation-induced forces: body is forced to oscillate with
wave excitation frequency and there are no incident waves
F Added mass: M
A
C
A
()
F Potential damping: D
P
()
F Weight and buoyancy: f
R
()
F Other damping eects, including skin friction, wave drift
damping, and vortex shedding damping
2. Froude-Krilo and Diraction Forces
3. Environmental forces: currents, waves, wind
4. Propulsion forces: thruster/propellor, control surfaces
Horizontal Motion
The state vector is
x =
h
v
T

T
r
T

T
i
T
= [u v w p q r x y z ]
T
For horizontal motion, the ship has zero heave, roll, and pitch;
the motion is restricted to surge, sway, and yaw
The control for horizontal motion typically involves thrusters
and rudders; linearized equations are:
(mX
u
) u = X
u
u + X
0
(mY
v
) v + (mx
G
Y
r
) r = Y
v
v + (Y
r
mU)r + Y
0
(mx
G
N
v
) v + (I
zz
N
r
) r = N
v
v (N
r
mx
G
U)r + N
0
The various X
u
etc. terms are constants; the X
0
etc. terms are
the control forces and moments
Horizontal Motion (2)
Linearized equations for horizontal motion are:
(mX
u
) u = X
u
u + X
0
(mY
v
) v + (mx
G
Y
r
) r = Y
v
v + (Y
r
mU)r +Y
0
(mx
G
N
v
) v + (I
zz
N
r
) r = N
v
v (N
r
mx
G
U)r +N
0
Clearly surge (u) is decoupled from sway (v) and yaw
()
The coupled sway-yaw system is

mY
v
mx
G
Y
r
mx
G
N
v
I
zz
N
r

x =

Y
v
Y
r
mU
N
v
N
r
mx
G
U

x +f
M x = A
0
x +f
x = M
1
A
0
x +M
1
f
x = Ax +Bu
Horizontal Motion (3)
The coupled sway-yaw system is

mY
v
mx
G
Y
r
mx
G
N
v
I
zz
N
r

x =

Y
v
Y
r
mU
N
v
N
r
mx
G
U

x +f
M x = A
0
x +f
x = M
1
A
0
x +M
1
f
x = Ax +Bu
Stability analysis leads to a simplied stability condition:
C = Y
v
N
r
+ N
v
(mUY
r
) > 0
The term C is called the vessel stability parameter
Ship design criteria are inuenced by making C > 0 true; for
example, adding more aft surface area drives N
v
positive, in-
creasing stablity
Stability Analysis via Routh-Hurwitz
Suppose we know A in terms of system parameters; i.e., A =
A(m, I, x

, u

, )
How can we determine stability using system parameters?
Routh-Hurwitz stability criteria
Develop the characteristic polynomial for A:
p() =
n
+ a
n1

n1
+ + a
0
Develop the Hurwitz matrix
H =

a
n1
a
n3
0
1 a
n2
0
0 a
n1
0
.
.
.
.
.
.
.
.
.
.
.
.
0 0 a
0

Stability Analysis via Routh-Hurwitz (2)

The Hurwitz matrix is
H =

a
n1
a
n3
0
1 a
n2
0
0 a
n1
0
.
.
.
.
.
.
.
.
.
.
.
.
0 0 a
0

Dene the principal minors of H by

1
= a
n1

2
= det

a
n1
a
n3
1 a
n2

.
.
.

n
= det H
Stability Analysis via Routh-Hurwitz (3)
Necessary and sucient conditions for the asymptotic stability
of x = Ax are

i
> 0 i = 1, , n
Since A depends on the parameters, p, the coecients of p()
depend on p, and the principal minors depend on p
Thus these conditions dene regions in parameter space where
the linearized system is asymptotically stable
Stable and unstable regions are separated by stability bound-
aries, which correspond to one or more eigenvalues crossing the
imaginary axis
If a real eigenvalue crosses the imaginary axis, then at that point
in parameter space a
0
(p) = 0 (exercise: convince yourself that
this statement must be true)
Stability Analysis via Routh-Hurwitz (4)
If a real eigenvalue crosses the imaginary axis, then at that point
in parameter space a
0
(p) = 0
If a c.c. pair crosses the imaginary axis, then

n1
(p) = 0
Thus the stability boundaries in parameter space are dened by
one of the two conditions :
a
0
(p) = 0

n1
(p) = 0
Stability Analysis via Routh-Hurwitz (5)
Suppose A =

A
11
A
12
A
21
A
22

. Then the characteristic polyno-

mial is p() =
2
(A
11
+ A
22
) + A
11
A
22
A
12
A
21
The principal minors of H are

1
= a
1
= (A
11
+ A
22
) > 0 A
11
+A
22
< 0

2
= det

a
n1
a
n3
1 a
n2

= (A
11
+A
22
)(A
11
A
22
A
12
A
21
) > 0
(A
11
+ A
22
)(A
11
A
22
A
12
A
21
) < 0
(A
11
A
22
A
12
A
21
) > 0
These conditions were used in the sway-yaw example. Details
are in 4.2 of Triantafyllou and Hover
Stability Analysis via Routh-Hurwitz (6)
Suppose A =

1 A
12
2 4

. Then the conditions are

1
> 0 1 4 < 0 X

2
> 0 (4 + 2A
12
) > 0
2A
12
> 4 A
12
> 2
Exercise: verify numerically that the appropriate stability
boundary condition is satised when A
12
= 2 and that the
stability condition developed here is valid