Chapter 4 Kinematics of Multibody Systems

4.1 Introduction

A multibody system is a collection of many bodies as shown in Figure 4.1.1. Adjacent

bodies may be connected in many ways with three translated and three rotational
degrees of freedom allowed between each pair of adjoining bodies. That is, the bodies
themselves may be either connected to each other or free to translate relative to each
other. The bodies may be either rigid or flexible. They may or may not form closed
loops. If a multibody system is composed entirely of connected bodies without closed
loops, it is called an open-chain or open-tree systems. Figure 4.1.1 shows such a
system. A closed loops is shown in Figure 4.1.2. In this book we will focus our
attention upon open-chain systems with connected rigid bodies.

The fundamental problem in studying multibody systems is organizing and describing

the complex geometry of the systems. Hence, to analyze the kinematics and dynamics
of such systems, a method developed by Huston et al. [1-3] will be used. This method
provides a useful accounting procedure from which the kinematics of any multibody
system is readily developed.

1
Figure 4.1.1 An open-chain multibody system

Figure 4.1.2 A closed loops multibody system

2
4.2 Geometrical Connection Configuration

Consider a multibody system as shown in Figure (4.1.1). Let the bodies of the system
be numbered as follows: First, select one of the bodies of the system as a reference
body, and call it Body 1, or (B 1). Next, number the bodies in ascending progression
away from Body 1 through the branches of the multibody system until all the bodies
are numbered. (see Figure (4.2.1))

When a multibody system is numbered (or labeled) as in Figure (4.2.1), each

body has a unique, adjacent, lower numbered body. Thus, a listing of the lower
numbered bodies for each body is developed. The lower numbered body for body 1 is
0 ( zero ) which refers to an inertial reference frame R in which the system moves.
The array containing these lower body numbers for each body of the system is called
"the Lower Body Array L(k)", which represents the list of body array. k is the index
corresponding to the body number. That is, B k . Thus, L(k) defines the connection
order between the bodies where L0 (k )  k , L1 ( k )  L(k ) , L2 ( k )  L( L( k )) ,
L3 (k )  L( L((L( k ))) , … etc. In this sense L(k) could be considered as an operator
on k.

For the system shown in Figure (4.2.1), L(k) is:

L0 ( k ) : (1, 2, 3, 4, 5, 6, 7, 8, 9, 10)
L1 ( k ) : (0, 1, 2, 1, 1, 5, 2, 7, 8, 9) (4.2.1)

The lower body array, L(k), is unique, and is a numerical description of the
geometrical connection configuration of the multibody system. That is, the multibody
system structure or topology (as in Figure 4.2.1) and the lower body array (as in
Equation 4.2.1) are equivalent.

Using the L( k ) , we can find its lower body array. Proceeding in this fashion, we can
continue to develop lower body arrays until all zeros occur. Using the multibody
system of Figure 4.2.1 as an example, we obtain the listing of Table 4.2.1 illustrating
the formation of the tree array by operator L.

3
To see how these lower body arrays might be useful in kinematic analyses, consider
body B10 of the system of Figure 4.2.1. We want to obtain the angular velocity of
B10 in R we could use the relative angular velocities and the addition formula, which
leads to the expression:

      
R
 B10  B9  B10  B8  B8  B7  B8  B2  B7  B1  B2  R  B1 (4.2.2)

or

      
  9  8  7  2  1
10  10 (4.2.3)

   
where R
 Bk   k is the angular velocity of B k relative to R and Bj
 Bk  k is the

angular velocity of B k relative to B j .

Example 4.2.1

Human body models have been used extensively in crash-victim simulation. In crash-
victim simulation, the human body model is placed within a vehicle model. An
acceleration ( or deceleration ) is given to the human body to simulate a vehicle crash
condition. Consider the human body model shows in Figure 4.2.2. It composed of 15
segments interconnected through revolute or ball-and-socket ( spherical ) joints.The
model is considered as a multibody system with a open chain configuration. We may
use the geometrical Connection Configuration, lower body arrays numbering scheme
to describe the configuration of the system. Table 4.2.2 givens as explicit lower body
arrays of the human body model.

4
 Table 4.2.1 Lower Body Arrays for the Multibody System of Figure 4.2.1

L0 (k ) 1 2 3 4 5 6 7 8 9 10
1
L (k ) 0 1 2 1 1 5 2 7 8 9
L2 ( k ) 0 1 0 0 1 1 2 7 8
3
L (k ) 0 0 0 1 2 7
L4 ( k ) 0 1 2
5
L (k) 0 1
6
L (k ) 0

Table 4.2.2 Lower Body Arrays for the Human Body Model of Figure 4.2.2
L0 (k ) 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
1
L ( k )1 2 3 4 5 4 7 4 9 2 11 12 2 14 15
2
L (k ) 0 1 2 3 4 3 4 3 4 1 2 11 1 2 14
L3 ( k ) 0 1 2 3 2 3 2 3 0 1 2 0 1 2
4
L (k ) 0 1 2 1 2 1 2 0 1 0 1
5
L (k) 0 1 0 1 0 1 0 0
L6 (k ) 0 0 0

5
Figure 4.2.1 Numbering for the Multibody System in Figure 4.1.1

6
Figure 4.2.2 Human Body Model

7
4.3 Coordinates

4.3.1 Generalized Coordinates

The number of degrees of freedom is the number of independent variables or

parameters (for example, translation variables or orientation angles) necessary to
specify the configuration of a system. For complex systems, such as a multibody
system, it is convenient to use a set of independent variables, q1 , q 2 , …, q r , called
generalized coordinates for these descriptive parameters. The choice of coordinates
however is not unique. We can use “relative coordinates”, “absolute coordinates”. On
various combinations of coordinates (parameters), relative coordinates are “natural”
coordinates of the system. They are useful for intuitive descriptions of the
configuration of the system. For example, in a human body model, it is convenient to
describe the head configuration in terms of angles at the neck, upper torso – instead of
using orientation angles in space. The absolute coordinates are the translation
variables and orientation angles measured in an inertia reference frame. Absolute
coordinates may be tedious in describing the configuration of a multibody system. In
this book we will exclusively use relative coordinates (that is, translations and
rotations between adjoining bodies) as our fundamental geometric variables.

Consider a multibody system containing N bodies. It can have as many as 6N degrees

of freedom: 3N translations and 3N rotations. Let these degrees of freedom be
characterized by 6N variables x r (r = 1, 2, ..., 6N), called "generalized coordinates".
These generalized coordinates can be arranged in 2N sets of 3 each, which depict the
translation and rotation of each body. Further it is convenient to arrange and number
the translation and rotation sets separately.

For the multibody system in Figure (4.3.1), if both translation and rotation are permitted
between each of the adjoining bodies there are 60 degrees of freedom. Let the 6N x r be
arranged as in Table (4.3.1). The translation and rotation classification of generalized
coordinates of the system are listed in Table (4.3.2).

Let the first 3N triplets of the coordinates describe the relative translation between the
adjoining bodies and let the remaining 3N triplets coordinates describe the relative
orientation between the adjoining bodies. Then, the coordinates are called body

8
coordinates. As an illustration Table (4.3.3) and Table (4.3.4) show the translation and
rotation classification of body coordinates of the system in Figure (4.3.1).
If Euler parameters are used, there are four rotational coordinates describing three
rotational degrees of freedom at each joint. Since these parameters are not
independent, there is a additional constraint equation as each joint as follow:

12   22   32   24  1 (4.3.1)

Consider the multibody system in Figure 4.3.1 again, which has 6N degrees of freedom. With
the use of Euler parameters there are 7N degrees of freedom, three translation coordinates and
four rotational coordinates for each body. As an illustration the generalized coordinates of
the system are listed in Table (4.3.5) and Table (4.3.6). Table (4.3.7) and Table (4.3.8)
show the body coordinates using Euler parameters for the system in Figure (4.3.1).

Total Coordinates: x r r = 1, 2, …, 6N

Table 4.3.1 Generalized coordinates of the System of Figure 4.3.1

Body Translation Rotation

B1 x1 x 2 x 3 x4 x5 x6
B2 x7 x8 x9 x 10 x 11 x 12
B3 x 13 x 14 x 15 x 16 x 17 x 18
. . . . . . .
. . . . . . .
. . . . . . .

x 6 ( N 1) 1 x 6 ( N 1)  2 x 6 ( N 1)  4 x 6 ( N 1)  5 x 6 ( N 1)  6

BN x 6 ( N 1)  3

Table 4.3.2 Translation and rotation of generalized coordinates of the system of Figure
4.3.1

9
Translation Coordinates: x km k = 1, 2, …, N; m = 1, 2, 3
Rotation Coordinates: x km k = 1, 2, …, N; m = 4, 5, 6

Table 4.3.3 Body Coordinates for the system of Figure 4.3.1

Body Translation Rotation

B1 x 11 x 12 x 13 x 14 x 15 x 16
B2 x 21 x 22 x 23 x 24 x 25 x 26
B3 x 31 x 32 x 33 x 34 x 35 x 36
. . . . . . .
. . . . . . .
. . . . . . .

x N1 x N 2 x N 3 x N 4 x N5 x N 6
BN

Table 4.3.4 Translation and Rotation Body Coordinates for the system of Figure 4.3.1

10
Total Coordinates: x r r = 1, 2, …, 7N

Table 4.3.5 Generalized coordinates using Euler Parameters of the System of Figure
4.3.1
Body Translation Rotation
B1 x1 x 2 x 3 x4 x5 x6 x7
B2 x 8 x 9 x 10 x 11 x 12 x 13 x 14
B3 x 15 x 16 x 17 x 18 x 19 x 20 x 21
. . . . . . .
. . . . . . .
. . . . . . .

x 7 ( N 1) 1 x 7 ( N 1)  2 x 7 ( N 1)  4 x 7 ( N 1)  5 x 7 ( N 1)  6

BN x 7 ( N 1)  3 x 7 ( N 1)  7

Table 4.3.6 Translation and Rotation of generalized coordinates using Euler

Parameters for the System of Figure 4.3.1

11
Translation Coordinates: x km k = 1, 2, …, N; m = 1, 2, 3
Rotation Coordinates: x km k = 1, 2, …, N; m = 4, 5, 6, 7
Constraint Equations: k = 1, 2, …, N
 
2
k1
2
k2  2
k3  2
k4 1

Table 4.3.7 Body Coordinates Using Euler Parameters of the system of Figure 4.3.1

Body Translation Rotation

B1 x 11 x 12 x 13 x 14 x 15 x 16 x 17
B2 x 21 x 22 x 23 x 24 x 25 x 26 x 27
B3 x 31 x 32 x 33 x 34 x 35 x 36 x 37
. . . . . . .
. . . . . . .
. . . . . . .

x N1 x N 2 x N 3 x N 4 x N5 x N 6 x N 7
BN

Table 4.3.8 Translation and Rotation of Body Coordinates Using Euler Parameters for
the system of Figure 4.3.1

12
Figure 4.3.1 A multibody system

13
4.3.2 Generalized Coordinates Derivatives

For a multibody system, the configuration can be defined by generalized coordinates.

However, the kinematics of the system is defined by the coordinate derivatives.

Consider again the multibody system in Figure (4.3.1). If we use generalized

coordinate derivatives to describe the kinematics of the system, then the coordinate
derivative, y r , would appear as listed in Table (4.3.9). If we choose relative velocity
and relative angular velocity components as coordinate derivatives, then the
coordinate derivatives, y r , would appear as listed in Table (4.3.10), where the r
are the components of the angular velocity of B k relative to its adjoining lower
numbered body. Similarly, the coordinate derivatives of the body coordinate for the
system of Figure 4.3.1 would appear as liasted in Table (4.3..11), where the ki are
the components of the angular velocity of Bk relative to its adjoining lower numbered
body.

In the sequel, these relative velocity and relative angular velocity will be used as
basic parameters to formulate the equations of motion of the multibody system.
Indeed, when the relative angular velocity components are used as generalized
coordinate derivatives, the "coordinates" themselves do not exist in explicit form.
That is, there are no single parameters which can be differentiated to produce the
angular velocity components. Hence, they are called "quasi-coordinates".

Consider again the two adjoining bodies B k and Bj as shown in Figure (4.3.2).

The relative displacement of these bodies may be measured by vector s :

   
s k  s k1 n j1  s k 2 n j2  s k 3 n j3 (4.3.2)


where n jm (m = 1, 2, 3) is the unit vector sets which is fixed in B j. Then, the relative

velocity v k of Bk with respect to Bj can be expressed as:

   
v k  s k1 n j1  s k 2 n j2  s k 3 n j3 (4.3.3)

14
In the same way, the relative angular velocity of B k with respect to Bj may be
expressed as:

   
 k   k1 n j1   k 2 n j2   k 3 n j3 (4.3.4)

If the generalized coordinate derivatives y r ( r = 1, 2, ..., 6N ) are used in Equation

(4.3.3) and (4.3.4), then the "generalized angular speeds" are

y r   km , ( k = 1, 2, ..., N; m = 1, 2, 3; r = 3(k-1) + m )
(4.3.5)

and the "generalized linear speeds" are

y r  s km , ( k = 1, 2, ..., N; m = 1, 2, 3; r = 3N + 3(k-1) + m ) (4.3.6)

15
 Body Translation Rotation
Derivatives Derivatives
B1 y1  x 1 , y 2  x 2 , y 3  x 3 y 4  x 4 , y 5  x 5 , y 6  x 6
B2 y 7  x 7 , y 8  x 8 , y 9  x 9 y10  x 10 , y11  x 11 , y12  x 12
. . . . . . .
. . . . . . .
. . . . . . .

y 6 ( N 1) 1  x
 6 ( N 1) 1 y 6 ( N 1)  4  x
 6 ( N 1)  4
BN y 6 ( N 1)  2  x
 6 ( N 1)  2 y 6 ( N 1) 5  x
 6 ( N 1) 5
y 6 ( N 1) 3  x
 6 ( N 1) 3 y 6 ( N 1)  6  x
 6 ( N 1)  6

Table 4.3.9 Generalized Coordinate Derivatives of the multibody System of Figure

4.3.1

16
 Body Translation Rotation
Derivatives Derivatives
B1 y1  x 1 , y 2  x 2 , y 3  x 3 y 4  4 , y 5  5 , y 6  6
B2 y 7  x 7 , y 8  x 8 , y 9  x 9  , y11  11
y10  10  , y12  12 
. . . . . . .
. . . . . . .
. . . . . . .

y 6 ( N 1) 1  x
 6 ( N 1) 1
y 6 ( N 1)  4  6 ( N 1)  4
BN y 6 ( N 1)  2  x
 6 ( N 1)  2
y 6 ( N 1) 3  x
 6 ( N 1) 3
y 6 ( N 1)  5  6 ( N 1) 5

y 6 ( N 1)  6  6 ( N 1)  6

Table 4.3.10 Generalized Coordinate Derivatives (Using

Relative Angular Velocity Components) of the
Multibody System of Figure 4.3.1

17
 Body Translation Rotation
Derivatives Derivatives
B1 y11  x 11 y14  x 14
y12  x 12 y15  x 15
y13  x 13 y16  x 16
B2 y 21  x 21 y 24  x 24
y 22  x 22 y 25  x 25
y 23  x 23 y 26  x 26
. . . . . . .
. . . . . . .
. . . . . . .

y N1  x
 N1 y N 4  x N 4
BN y N 2  x N 2 y N 5  x N 5
y N 3  x N 3 y N 6  x N 6

Table 4.3.11 Coordinate derivatives of the body coordinate for the multibody System
of Figure 4.3.1

18
 Body Translation Rotation
Derivatives Derivatives
B1 y11  x 11 
y14  14
y12  x 12 
y15  15
y13  x 13 
y16  16
B2 y 21  x 21 y 24  24
y 22  x 22 y 25  25
y 23  x 23 y 26  26
. . . . . . .
. . . . . . .
. . . . . . .

y N1  x
 N1 y N 4  N 4
BN y N 2  x N 2 y N 5  N 5
y N 3  x N 3 y N 6  N 6

Table 4.3.12 Coordinate Derivatives of the body

coordinate of the Multibody System of Figure
4.3.1(Using Relative Angular Velocity
Components)

19
Figure 4.3.2 Two typical adjoining bodies of the system

20
4.4 Angular Velocities

Consider a typical body Bk ( k=1,2,...,N ) of a multibody system, the angular velocity


of Bk relative to the inertial frame R ( absolute angular velocity ),  k , may be
obtained from the relative angular velocities by using the addition formula for angular
velocities [1] as follows:

   
 k  1  2      k (4.4.1)

where the terms on the right side with a prime, over the symbol represent the angular
velocity of a body relative to its adjacent lower numbered body. The sum is taken
between the bodies of B1 and Bk. For example, the angular velocity of B10 in Figure

(4.3.1), 10 is

      
10  1  2  7  8  9  10
 (4.4.2)

where


1 is the angular velocity of B1 relative to R

2 is the angular velocity of B2 relative to B1

7 is the angular velocity of B7 relative to B2

8 is the angular velocity of B8 relative to B7

9 is the angular velocity of B9 relative to B8

 is the angular velocity of B10 relative to B9
10 (4.4.3)

Using the lower body array concept [2], it is seen that

L0 (10)  10
L1 (10)  L( L0 (10))  L(10)  9
L2 (10)  L( L1 (10))  L(9)  8
L3 (10)  L(L2 (10))  L(8)  7
L4 (10)  L(L3 (10))  L(7)  2
L5 (10)  L( L4 (10))  L( 2)  1 (4.4.4)


Therefore, 10 can be expressed as

21
 5
 
10   u where u = Lt (10) (4.4.5)
t 0

Hence, the absolute angular velocity of a typical body B k can be expressed as

v
 
 k   u where u = Lt (k) (4.4.6)
t 0

where v is the index such that L v (k )  1 .

Therefore, by using lower body arrays and relative angular velocity concepts, the
absolute angular velocity of each body of a multibody system can be calculated.

Consider again the two typical adjoining bodies of a multibody system B j and B k
 
as shown in Figure (4.3.2). Let mutually perpendicular unit vectors n ji and n ki (i =
 
1, 2, 3) are fixed in B j and B k . Let n ji and n ki be originally aligned with each
other and then the orientation of B k relative to B j may be defined by the relative
 
inclinations of the n ji and n ki or three rotation angles  k ,  k , and  k . The
 
configuration graphs for the relation between the n ji and the n ki are shown in
 
Figure (4.4.1). Then n ji and n ki can be expressed in the form:

 
[ n j ]  [SJK ][ n k ] (4.4.7)

or, in index notation as:

 
n jm  SJK mn n kn (4.4.8)

where

[SJK ]  [JK ][ JK ][ JK ] (4.4.9)

where

1 0 0 

[JK ]  0 cos  k  sin  k  (4.4.10)
0 sin  k cos  k 

22
 cos  k 0 sin  k 
[ JK ]   0 1 0  (4.4.11)
 0 0 cos  k 

cos  k  sin  k 0
[ JK ]   sin  k cos  k 0 (4.4.12)
 0 0 1


The relative angular velocity k of body B k with respect to B j , may be written as
(see Figure (4.4.1)) [2]

   
 k n j1   k n j2   k n k 3
k   (4.4.13)

or

   
k  1k n j1  k 2 n j2  k 3 n k 3 (4.4.14)

23
 
In terms of n jn (n=1,2,3)

 
n j1   n1 n jn
 
n j2  JK n 2 n jn (4.4.15)
 
n k 3  JK np  JK p 3 n jn


Hence k [2] becomes

  JK   JK  JK )n

k  (
 k  n1   k n2 k np p3 jn (4.4.16)

where  nm are the components of Kronecker's delta function:

1 nm
 nm   (4.4.17)
0 nm


Let n om (m = 1, 2, 3) be mutually perpendicular unit vectors fixed in R (inertial

reference frame). Then the n jn may be expressed as:

 
n jn  SOJ mn n om (4.4.18)

 
Hence, k may be expressed in terms of the unit vectors, n om , fixed in R as:

 
 k  n1   k JK n 2   k JK np  JK p 3 )n om
k  SOJ mn ( (4.4.19)

By substituting from Equation (4.4.19) into Equation (4.4.5), we can express the
angular velocity of a typical body of a multibody system in a linear form:

 
 k   km x r n om (k  1,..., N; r  1,...,3N; m  1,2,3)
(4.4.20)
.
where the x  r represent the orientation angle derivatives and the coefficients  krm ,
are called "partial angular velocity components".
It is seen that the  krm are:

24
  S0J m1

 hrm   S0J mn JK n 2 (3k  3  r  3k )
S0J JK  JK
 mn np p3

(4.4.21)

where j = L(k) (4.4.22)

and
 krm   jrm ( r  3k-3) (4.4.23)
finally,

 krm = 0 ( r > 3k) (4.4.24)

If the generalized coordinates derivative are used, Equation (4.4.14) may be written in
the form:

   
k  y h 1 n j1  y h  2 n j2  y h 3 n j3 (4.4.25)

where h=3(K-1)

or

k  y h  m n jm (4.4.26)


In terms of the unit vectors in inertial frame R, k can be written as


k  y h  p S0J mp n om (4.4.27)


Then, the partial velocity  k /  y h  p is:

 
 k /  y h  p  SOJ mp n om (4.4.28)

Therefore, the angular velocity of B k in inertial frame, R, is

 
 k   krm y r n om (4.4.29)

25
for 3k-3 < r  3k

 krm  S0J mp (4.4.30)

where j = L(k), p = l - (3k-3)

for r  3k-3

 krm   jrm (4.4.31)

for r > 3k

 krm  0 (4.4.32)

For the example multibody system of Figure 4.4.2, the  krm is shown in Table
(4.4.1).

26
Table 4.4.1  krm for the multibody system of Figure 4.4.2

r 1 2 3 4 5 6 7 8 9 10
k
1 2 3 I I I I I I I I I I
4 5 6 0 S01 S01 0 0 0 S01 S01 S01 S01
7 8 9 0 0 S02 0 0 0 0 0 0 0
10 11 12 0 0 0 S01 0 S01 0 0 0 0
13 14 15 0 0 0 0 S01 0 0 0 0 0
16 17 18 0 0 0 0 0 S05 0 0 0 0
19 20 21 0 0 0 0 0 0 S02 S02 S02 S02
22 23 24 0 0 0 0 0 0 0 S07 S07 S07
25 26 27 0 0 0 0 0 0 0 0 S08 S08
28 29 30 0 0 0 0 0 0 0 0 0 S09

27
    
i n ji n ji n ki n ki
1
k
2
k
3
k

Figure 4.4.1 Configuration graph for relative orientation angles

Figure 4.4.2 A multibody system

28
4.5 Angular Accelerations

By differentiating the angular velocity in Equation (4.4.29), we obtain the angular

acceleration of body Bk in R as:

  
 k  d ( k ) / dt  ( km 
 r   krm y r )n om (4.5.1)

.
The 
 krm can be obtained by differentiating Equations (4.4.30), (4.4.31), (4.4.32):

for 3k-3 < r  3k

 J
 krm  S0 (4.5.2)
mp

where j = L(k), p = r - (3k-3)

for r  3k-3


 krm  
 jrm (4.5.3)

for r > 3K


 krm  0 (4.5.4)

To continue the example multibody system of Figure 4.4.2, the 

 krm is shown in
Table (4.5.1).

29
Table 4.5.1 
 krm for the multibody system of Figure 4.4.2

r 1 2 3 4 5 6 7 8 9 10
k
1 2 3 0 0 0 0 0 0 0 0 0 0
4 5 6 0  1 S0
S0  1 0 0 0  1 S0
S0  1 S0
 1 S0
 1

7 8 9 0 0  2
S0 0 0 0 0 0 0 0
10 11 12 0 0 0  1
S0 0  1
S0 0 0 0 0
13 14 15 0 0 0 0  1
S0 0 0 0 0 0
16 17 18 0 0 0 0 0  5
S0 0 0 0 0
19 20 21 0 0 0 0 0 0  2 S0
S0  2 S0
 2 S0
 2

22 23 24 0 0 0 0 0 0 0  7 S0
S0  7 S0
 7

25 26 27 0 0 0 0 0 0 0 0  8 S0
S0  8

28 29 30 0 0 0 0 0 0 0 0 0  9
S0

30
4.6 Mass Center Positions

Consider the two typical bodies Bj and Bk shown in Figure (4.6.1). Let Oj and Ok, be
the local reference points in Bj and Bk, respectively. Let Qk fixed in Bj, be the local

observation point of Bk. Let s k represent the relative position between Bk and Bj.
   
When s k  0 , Qk and Ok are coincident. Let q k locate Qk relative to Oj and let rk
locate Gk (the mass center of Bk) relative to Ok.


The position vector p k which locates the mass center Gk of Bk relative to the
origin point O, of the inertial reference R, may be expressed as:

     
p k  q 1  s1  ...  q k  s k  rk (4.6.1)

or

v
   
p k   (q u  s u )  rk (4.6.2)
t 0

where u = Lt (k), v is the index such that Lv(k) = 1

 
Let unit vectors n om (m = 1, 2, 3) be fixed in inertial reference frame R. Let n jm and

n km (m = 1, 2, 3) be mutually perpendicular unit vectors of the local coordinate fixed
  
in frame Bj and Bk, respectively. Thus vectors q k , rk and s k may be expressed as:

  
q k  q km n jm  SOJ mn q jn n om (4.6.3)

  
rk  rkm n km  SOK mn rkn n om (4.6.4)

 
s k  SOJ mn s kn n om (4.6.5)

Therefore by substituting from Equation (4.6.3), (4.6.4) and (4.6.5) Equation (4.6.2)
may be expressed as:

v
  
p k   (SOSmn (q sn  s sn )n om  SOK mn rkn n om (4.6.6)
t 0

31
where

s = Lt+1 (k) and Lv (k) = 1 (4.6.7)

32
Figure 4.6.1 Two Adjoining Bodies

33
4.7 Mass Center Velocities

The velocity of the mass center of Bk and Gk, can be obtained by differentiating
Equation (4.6.6) with respect to time in inertial reference R as:


 dp Rk
vk  (4.7.1)
dt

or

 v  
vk  { ([S0Smn (qsn  ssn )  S0Smns sn ]  S0K mnrkn}n om
 (4.7.2)

t 0

or

 v 
vk  { ([eghmsrg yrS0Shn (qsn  ssn )  S0Smns sn ]  eghmkrg yrS0K hn rkn }n om
t 0

(4.7.3)

where

s = Lt+1 (k), and Lv (k) = 1 (4.7.4)

In Equation (4.7.3) the y r ( r = 1, ..., 3N ), the generalized coordinate derivatives,

can be defined for l from 3N+1 to 6N as:

y r  s sn (3N+1  r  6N) (4.7.5)

where

34
 r = 3(N+s-1) + n (4.7.6)


By substituting Equation (4.7.5) into Equation (4.7.3) we see that the v k may be
written as:

 
v k  v krm y r n om (4.7.7)

where v krm ( k = 1, 2, ..., N; r = 1, 2, ..., 6N; m = 1, 2, 3 ), called "partial velocity

components", are
for r  3N

v
v krm   ([e ghm srg S0S hn (q sn  s sn )]  e ghm  krg S0K kn rkn (4.7.8)
t 0

for 3(K+N)-3 < r  3(K+N)

v km  S0J mn , n = r - 3(N+s-1), j = L(k) (4.7.9)

for r  3(K+N) - 3

v krm  v jrm (4.7.10)

for r > 3(K+N)

v krm  0 (4.7.11)

If we use the relative angular velocity components as generalized coordinate

derivatives, the v krm for 3N < r  6N, are the same as the  krm for 0 < r  3N.
That is,

v krm   k ( r 3 N ) m r = 3N+1,...,6N (4.7.12)

35
4.8 Mass Center Accelerations

The acceleration of Gk in inertial reference frame R may be written as


 dv Rk
ak  (4.8.1)
dt

or

 
a k  ( v krm y r  v krm y r ) n om (4.8.2)

where v krm is given by Equation (4.7.8) to (4.7.11), and the v krm may be obtained
by differentiating Equations (4.7.8) to (4.7.11) with respect to time as:
for r  3N,

v
v krm   [e ghm 
 srg S0S hn (q sn  s sn )  e ghm srg S0 S hn (q sn s sn )  e ghm  rg S0S hn s sn ]
t 0

 e ghm rkn  K
 krg S0K hn  e ghm rkn  krg S0 (4.8.3)
hn

for 3(K+N)-3 < r  3(K+N)

v J
 krm  S0 (4.8.4)
mn

where j = L(k) (4.8.5)

for r  3(k+N)-3

 krm  v
v  jrm (4.8.6)

finally, for r > 3(k+N)

v krm  0 (4.8.7)

Again, if relative angular velocity components are used as generalized coordinate

derivatives, the form of the v krm for 3N < r  6N is the same as the  krm
for 0 < r  3N. That is,

 krm  
v  k ( r 3 N ) m (4.8.8)

36
where r = 3N+1,..., 6N (4.8.9)

Reference
[1] Huston, R.L., and C.E. Passerello, and M.W. Harlow, “Dynamics of Multirigid-
Body Systems,” Journal of Applied Mechanics, Vol. 45, 1978, pp. 889-894.
[2] Huston, R.L., and C.E. Passerello, “On Multi-Rigid-Body Systems Dynamics,”
Computers and Structures, Vol. 10, 1979, pp. 439-446.
[3] Huston,R.L., Multibody Dynamics, Butterworth-Heinemann, 1990.

37