Jacobians:

Velocities and Static Force

1
Introduction
In this Chapter the following concepts will be studied:
• The linear and angular velocity of a rigid body to analyze the
motion of the manipulator
• Forces that act on a rigid body (application on static forces of
a manipulator)
 Jacobian matrix relations

2
 Linear & Angular Velocities - Frames

• When describing the velocity (linear or angular) of an

object, there are two important frames that are being used:

– Represented Frame (Reference Frame) :

This is the frame used to represent (express) the
object’s velocity.

– Computed Frame
This is the frame in which the velocity is measured
(differentiate the position).
3
Frame - Velocity

• As with any vector, a velocity vector may be described in terms

of any frame, and this frame of reference is noted with a leading
superscript.
• A velocity vector computed in frame {B} and represented in
frame {A} would be written

Represented
(Reference Frame)

A
dB
( VQ ) 
A B
PQ
dt
Computed
(Measured)

4
Differentiation of position vectors

Derivative of a vector:

d B
Q (t  t )  B
Q(t ) Q
B
VQ  Q  lim
B

dt t 0 t

We are calculating the derivative of Q relative to frame B.

5
Differentiation of position vectors

A velocity vector may be described in terms of any frame:

 V   dt
A
A B dB
Q Q
We may write it:

 V 
A B
Q
A
B
B
R VQ .

Special case: Velocity of the origin of a frame relative to some

understood universe reference frame

6 vC  UVCORG
 Example 5.1:

100 mph

30 mph

U
dU
PCORG U VCORG  vC  30 Xˆ .
dt
( VTORG )C vT UCRvT UC R(100 Xˆ )CU R 1100 Xˆ .
C U

C T
( VCORG )TCRT VCORG UCRUT RT VCORG CU R 1UT R70 Xˆ .
7
 Angular velocity vector: 

Linear velocity  attribute of a point

Angular velocity  attribute of a body

Since we always attach a frame to a body we can

consider angular velocity as describing rotational
motion of a frame.

8
Angular velocity vector: 
A
 B describes the rotation of frame {B} relative to {A}
Direction of  B
A

indicates instantaneous axis

of rotation

Magnitude of B A

indicates speed of rotation

  B   the angular velocity of {B}

C A

relative to {A} expressed in {C}

In the universe frame {U} U C  C

9
 Linear Velocity - Rigid Body
• Given: Consider a frame {B} attached
to a rigid body whereas frame {A} is
fixed. The orientation of frame {A} with
respect to frame {B} is not changing as
a function of time BA R  0

• Problem: describe the motion of the

vector B PQ relative to frame {A}

• Solution: Frame {B} is located relative

to frame {A} by a position vector A PBORG
and the rotation matrix BA R (assume
that the orientation is not changing in
time BA R  0 ) expressing both
components of the velocity in terms of
frame {A} gives
A
VQ  AVBORG  A( BVQ ) AVBORG  BAR BVQ
10
 Angular Velocity - Rigid Body
 Given: Consider a frame {B} attached
to a rigid body whereas frame {A} is
fixed. The vector B PQ is constant as
view from frame {B} BVQ  0

 Problem: describe the velocity of the

PQ

vector B PQ representing the point Q

relative to frame {A}
B
VQ  0
 Solution: Even though the vector B PQ VBORG  0
A

is constant as view from frame {B} it

is clear that point Q will have a
velocity as seen from frame {A} due
11 to the rotational velocity 
A
B
Angular Velocity - Rigid Body

 The figure shows to instants

of time as the vector A PQ
rotates around A  B This is
what an observer in frame
{A} would observe. A
PQ sin 

 APQ
 The Magnitude of the
differential change is A
PQ sin 
A

  
PQ (t )
A
PQ (t  t )
 PQ 
A A
 B t A
PQ sin 
 Using a vector cross product
we get
 APQ
 AVQ  A  B  APQ
t
12
Angular Velocity - Rigid Body

In the general case, the vector Q may also be changing with respect to
the frame {B}. Adding this component we get.

VQ 
A
 V    P
A B
Q
A
B
A
Q

Using the rotation matrix to remove the dual-superscript, and since the
A
description of A
PQ at any instance is B R BPQ we get

VQ  BAR BVQ  A  B BAR BPQ

A

13
Simultaneous Linear and Rotational Velocity
 The final results for the derivative of a vector in a moving frame

(linear and rotation velocities) as seen from a stationary frame

 Vector Form B
PQ
A
B

A
VQ  AVBORG  BAR BVQ  AB BAR BPQ

 Matrix Form

A
VQ  AVBORG  BAR BVQ  BAR BAR BPQ 

14
Definitions - Angular Velocity
 Angular Velocity: The instantaneous rate of change in the
orientation of one frame relative to another.

Point - 1D
Linear Velocity

Plane - 2D / Body - 3D

Angular Velocity Plane - 2D / Body - 3D

15
Definitions - Angular Velocity
 Just as there are many ways to represent orientation (Euler Angles, Roll-Pitch-

Yaw Angles, Rotation Matrices, etc.) there are also many ways to represent the
rate of change in orientation.

Angular Velocity
Representation

Angular Velocity Angular Velocity

Vector Matrix

 The angular velocity vector is convenient to use because it has an easy to grasp

physical meaning. However, the matrix form is useful when performing

algebraic manipulations.
16
 Definitions - Angular Velocity - Vector
 Angular Velocity Vector: A vector whose direction is the
instantaneous axis of rotation of one frame relative to another and
whose magnitude is the rate of rotation about that axis.

 x 
A
 B   y 
 
  z 

 The angular velocity vector

A
 B describes the instantaneous change
of rotation of frame {B} relative to frame {A}
17
 Definitions - Angular Velocity - Matrix
A
 The rotation matrix ( B R ) defines the orientation of frame {B} relative to frame {A}.

Specifically, the columns of BA R are the unit vectors of {B} represented in {A}.

A
 
A
B R     P  
 BP
 x
B
y
B
Pz 

 

 If we look at the derivative of the rotation matrix, the columns will be the velocity of

each unit vector of {B} relative to {A}:

A
 
A 
B R 
d A
dt
B     V   
R   BV
 x
B
y
B
Vz 

 

18
 Definitions - Angular Velocity - Matrix
A
 The relationship between the rotation matrix B R and the derivative of
the rotation matrix A
B R can be expressed as follows:
A
B R  BAR BA R
A A
   
  V   
 BV
 x
B
y
B
Vz   BAR 

   P  
 BP
 x
B
y
B
Pz 

   

 where
A
B R  is defined as the angular velocity matrix

 0  z y   x 
 
 B   y 
A 
R
B    z 0  x  A
 
  y x 0    z 
19
Angular Velocity - Matrix & Vector Forms
Matrix Form Vector Form

 0  z y   x 
 
Definition A 
R   z 0  x  A
 B   y 
B 
 
  y x 0    z 

Multiply by Constant k 
A
B R  k 
A
B

 x  x
Multiply by Vector  A   
B R  y 
A
B   y
 
 z   z 

Multiply by Matrix  R 
s
t R
A
B   R 
s
t
T
 R 
s
t
A
B 
20
 Simultaneous Linear and Rotational Velocity -
Vector Versus Matrix Representation
Vector Form Matrix Form

A
VQ  AVBORG  BAR BVQ  AB BAR BPQ A
VQ  AVBORG  BAR BVQ  BAR BAR BPQ 

i j k
  P  x y  z  i ( y Pz   z Py )  j ( x Pz   z Px )  k ( x Py   y Px )
Px Py Pz

 0  z  y   Px     z Py   y Pz 
   
R  P    z 0   x   Py     z Px   x Pz 
  y x 0   Pz    y Px   x Py 


21
Changing Frame of Representation - Linear Velocity

 Recall that
 


 R
A
P 
A 

BT 
A B B org

 
 
0 0 0 1 
 so that the derivative is

     
A  d

  R
A
P 
A  
  R 
A
 P 
A  
    A
R  BA R  V 
A 

BT 
B B org B B org B B org

dt      
     
0 0 0 1  0 0 0 0  0 0 0 0 

22
Summary – Changing Frame of Representation
 Linear and Rotational Velocity
A
B

 Vector Form
A
VQ  AVBORG  BAR BVQ  AB BAR BPQ
B
PQ

 Matrix Form
A
VQ  AVBORG  BAR BVQ  BAR BAR BPQ 

 Angular Velocity

 Vector Form
A
C  A  B  BAR B C

 Matrix Form A
C R   BAR   BAR CBR  BA RT

23
 Position Propagation

• The homogeneous transform matrix provides a complete

description of the linear and angular position relationship
between adjacent links.
i 1
o
T  T T T
i
o
1
1
2 i

• These descriptions may be combined together to

describe the position of a link relative to the robot base
frame {0}.

24
 Motion of the Link of a Robot
• In considering the motion of a robot link we will always use link
frame {0} as the reference.

Where: vi- is the linear velocity of the origin of link frame (i)
with respect to frame {0} (Computed AND Represented)
 i - is the angular velocity of the origin of link frame (i)
with respect to frame {0} (Computed AND Represented)

25
 Velocities - Frame & Notation
Expressing the velocity of a frame {i} (associated with link i ) relative
to the robot base (frame {0}) using our previous notation is defined as

   
follows: 0
vi  0Vi  0Vi
       
0 0 0
i i i

The velocities differentiate (computed) relative to the base frame

{0} are often represented relative to other frames {k}. The
following notation is used for this conditions
k
   
k 0
vi  Vi  0kR 0Vi  0kR  vi
      R    R  i
k k 0 k 0 k
i i 0 i 0

26
 Velocity Propagation
 Given: A manipulator - A chain of rigid bodies each one capable of
moving relative to its neighbor
 Problem: Calculate the linear and angular velocities of the link of a
robot
 Solution (Concept): Due to the robot structure (serial mechanism)
we can compute the velocities of each link in order starting
from the base.
The velocity of link i+1 will be that of link i , plus whatever
new velocity components were added by joint i+1

27
 Velocity of Adjacent Links - Angular Velocity

 From the relationship developed previously

A
C  A  B  BAR B C
 we can re-assign link names to calculate the velocity of any
link i relative to the base frame {0}
 A0

 Bi
 C  i 1

0
i 1 0 i  0i Ri i 1

i 1
 By pre-multiplying both sides of the equation by R ,we
0
can convert the frame of reference for the base {0} to frame
28
{i+1}
 Velocity of Adjacent Links - Angular Velocity
i 1
0 R 0 i 1 i 01R 0 i i 01R 0i Ri i 1

 Using the recently defined notation, we have

i 1 i 1i  i i1R i i 1
i 1

i 1
i 1 - Angular velocity of frame {i+1} measured relative to the
robot base, and expressed in frame {i+1}
-
i 1
i - - Angular velocity of frame {i} measured relative to the robot
base, and expressed in frame {i+1}

i 1
i R i i 1 - Angular velocity of frame {i+1} measured relative to frame
{i} and expressed in frame {i+1}

29
 Velocity of Adjacent Links - Angular Velocity
i 1 i 1i  i i1R i i 1
i 1

• Angular velocity of frame {i} measured relative to the robot base,

expressed in frame {i+1}

i i i1Rii
i 1

30
 Velocity of Adjacent Links - Angular Velocity
i 1 i 1i  i i1R i i 1
i 1

• Angular velocity of frame {i+1} measured (differentiate) in

frame {i} and represented (expressed) in frame {i+1}
• Based on the frame attachment convention in which we assign
the Z axis pointing along the i+1 joint axis such that the two are
coincide (rotations of a link is preformed only along its Z- axis)
we can rewrite this term as follows:

 0 
i 1 i
R    0 
i+1 i i 1  
i i 1 

31
 Velocity of Adjacent Links - Angular Velocity
 The result is a recursive equation that shows the angular
velocity of one link in terms of the angular velocity of the
previous link plus the relative motion of the two links.
 0 
i 1
i 1 i i1Rii   0 
 
i 1 

 Since the term i 1 depends on all previous links through

i 1

this recursion, the angular velocity is said to propagate from

the base to subsequent links.

32
 Velocity of Adjacent Links – Linear Velocity
From the relationship developed previously (matrix form)
VQ  AVBORG  BAR BVQ  AB BAR BPQ
A

A
VQ  AVBORG  BAR BVQ  BAR BAR BPQ 
we re-assign link frames for adjacent links (i and i +1) with the velocity
computed relative to the robot base frame {0}
 A0

 Bi
 C  i 1

Vi 1 0i R 0i R iPi 1  0Vi  0i R iVi 1
0

i 1
 By pre-multiplying both sides of the equation by ,we can 0 R
convert the frame of reference for the left side to frame {i+1}
33
 Velocity of Adjacent Links – Linear Velocity

i 1
0 R 0Vi 1 i 01R 0i R 0i R iPi 1  i 01R 0Vi  i 01R 0i R iVi 1

 Which simplifies to
i 1
0 R 0Vi 1 i 01R 0i R 0i R iPi 1  i 01R 0Vi  i i1R iVi 1

i 1
 Factoring out i R from the left side of the first two terms

i 1
0 R 0Vi 1 i i1R 0i R 0i R 0i R iPi 1  0iR 0Vi  i i1R iVi 1

34
 Velocity of Adjacent Links – Linear Velocity
i 1
0 R 0Vi 1 i i1R 0i R 0i R  0i R iPi 1  0iR 0Vi  i i1R iVi 1
i 1
i R iVi 1- Linear velocity of frame {i+1} measured relative to frame {i}
and expressed in frame {i+1}

 Based on the frame attachment convention in which we assign

the Z axis pointing along the i+1 joint axis such that the two
are coincide (translation of a link is preformed only along its Z-
axis) we can rewrite this term as follows:

 0 
i 1
R iVi 1  0 
i  
di 1 

35
 Velocity of Adjacent Links – Linear Velocity

 0 
i 1 0
0 R Vi 1  i 1
i R  i 0 0 i
R R
0 i i R P 
i 1 0
i 0
R Vi    0 
 
di 1 
0
i
R 0i R 0i R 0iR 0i R 0i RT 0iR 0 i 0iRi ii

i 1
0 R 0Vi 1 i 1vi 1 0
i
R 0Vi  i vi

36
 Velocity of Adjacent Links – Linear Velocity

 The result is a recursive equation that shows the linear

velocity of one link in terms of the previous link plus the
relative motion of the two links.

 0 
i 1
 
vi 1 i i1R i i i Pi 1  i vi   0 
di 1 

 Since the term i 1

vi 1 depends on all previous links through
this recursion, the angular velocity is said to propagate from
the base to subsequent links.

37
 Velocity of Adjacent Links – Summary
 Angular Velocity
0 - Prismatic Joint

 0 
i 1
i 1 i i1Rii   0 
 
i 1 

 Linear Velocity
0 - Revolute Joint

 0 
i 1
vi 1  i i1R  ii  i Pi 1  i vi    0 
 di 1 

38
Angular and Linear Velocities – 2links Robot - Example

c1  s1 0 0
c 2  s 2 0 l1
 s1 c1 0  s2 c2
1T 
0 
0
  0 0 
2T 
1
0 0 1 0 0
  0 1 0
0 0 0 1  
0 0 0 1

1 0 0 l 2
0 1 0 0 
3T 
2 
0 0 1 0
 
0 0 0 1

39
Angular and Linear Velocities – 2links Robot - Example
 0 
i 1
i 1 i i1Rii   0 
i 1 
 For i=0
 0   c1 s1 0 0  0   0 
1
1 01R 00   0    s1 c1 0 0   0    0 
1   0 0 1 0 1  1 
 For i= 1

 0   c2 s 2 0  0   0   0 
2
2  12 R 11   0     s 2 c 2 0   0    0    0 
 2   0 0 1  1   2  1   2 

40
Angular and Linear Velocities – 2links Robot - Example
For i=2
1 0 0   0   0 
3
3  23 R 22  0  0 1 0   0    0 
0 0 1  1   2  1   2 

 0 
i 1
vi 1  i i1R  ii  i Pi 1  i vi    0 
 di 1 
For i=0

 c1 s1 0  0 0 0  0

v1  01R00 0 P1  0v0    s1 c1 0  0  0  0   0
1  
         
 0 0 1  0 0 0  0
41
Angular and Linear Velocities – 2links Robot - Example
For i=1
 c 2 s 2 0    0  l1 0  
 
2
v2  12 R  11  1P2  1v1    s 2 c 2 0    0    0   0  
 0 0 1   1   0  0  
 c 2 s 2 0   0  l1s 21 
 
   s 2 c 2 0  l11   l1c 21 
 0 0 1   0   0 

For i=2
1 0 0    0  l 2  l1s 21  
  
3
v3  23 R  22  2 P3  2 v2   0 1 0    0    0   l1c 21  
0 0 1   1   2   0   0  
 

42
Angular and Linear Velocities – 2links Robot - Example
 c1  s1 0  c 2  s 2 0  c1c 2  s1s 2 c1s 2  s1c2 0 
 c1 0  s 2 c 2 0   s1c 2  c1s 2  s1s 2  c1c 2 0 
2 R   s1
0
   
 0 0 1   0 0 1   0 0 1 
 c12  s12 0
  s12 c12 0  30 R
 
 0 0 1 

 c12  s12 0  l1s 21 

 
0 0 3 


 
v3  3 R  v3  s12 c12 0 l1c 21  l 2 1   2  
 0 0 1   0

 

43
  c i  s i 0 ai 1 
 s c c i ci 1  si 1  si 1d i 
i 1
iT 
 i i 1 
 s i si 1 c i si 1 ci 1 ci 1d 
 
 0 0 0 1 
c1  s1 0 0 c2  s 2 0 L1
 s1 c1 0 0 0 0 1 0 
1T 
0   2T 
1  
0 0 1 0  s2 c2 0 0
   
0 0 0 1 0 0 0 1
c3  s 3 0 L 2
 s3 c3 0 0
3T 
2  
0 0 1 0
 
0 0 0 1
i  i 1 ai 1 di i
1 0 0 0 1
2 90 L1 0 2
3 0 L2 0 3
4 0 L3 0 0

44
  From the DH parameter table, we can specify the homogeneous
transform matrix for each adjacent link pair:
 c i  s i 0 ai 1 
 s c c i ci 1  si 1  si 1d i 
i 1
iT 
 i i 1 
 s i si 1 c i si 1 ci 1 ci 1d 
 
 0 0 0 1 
c1  s1 0 0 c2  s 2 0 L1
 s1 c1 0 0 

0
 0  1 0
1T   
0
2T 
1
0 0 1 0  s2 c2 0 0
   
0 0 0 1
 0 0 0 1 

c3  s 3 0 L 2
 s3 c3 0 0
3T 
2  
0 0 1 0
 
0 0 0 1
45
 Compute the angular velocity of the end effector frame relative to the
base frame expressed at the end effector frame.

 0 
i 1
i 1 i i1Rii   0 
 
i 1 

 For i=0

 0   c1 s1 0 0  0   0 
1
1 01R 00   0    s1 c1 0 0   0    0 
        
1   0 0 1 0 1  1 

46
  For i=1
 0   c2 0 s 2  0   0   s 21 
 
2
2 12R11   0    s 2 0 c 2  0    0   c 21 
2   0  1 0  1  2   2 
 For i=2

 0   c3 s3 0  s 21   0   s 231 
          
3
3  2 R  2  0   s3 c3 0 c21   0   c231 
3 2
     
3   0 0 1  2  3  2  3 
 For i=3
0 1 0 0  s 231  0  s 231 
           
4
4 3 R 3  0  0 1 0  c231   0   c231 
4 3
     
 Note 0 0 0 1 2  3  0 2  3 

47 3  4 4
3
 Compute the linear velocity of the end effector frame relative to
the base frame expressed at the end effector frame.
 Note that the term involving the prismatic joint has been dropped
from the equation (it is equal to zero).

0
 0 
i 1
vi 1 i i1Ri i Pi 1 ivi    0 
 
di 1 

48
  For i=0

 c1 s1 0  0 0 0  0

v1  01R00 0 P1  0v0    s1 c1 0  0  0  0   0
1  
         
 0 0 1  0 0 0  0
 For i=1

 c2 0 s 2  0   L1 0   0 
v2 12R111P2 1v1   s 2 0 c 2  0    0   0    0 
2  

 0  1 0  1   0  0   L11 

49
  For i=3

 c3 s3 0  s 21   L 2  0  
      
3
v3  2 R 2  P3  v2    s3 c30 0 c 21    0    0  
3 2 2 2   
 0 0 1  2   0   L11  

 c3 s 3 0    0   L 2s32 
        
   s 3 c3 0    L 21    L 2c3 2 
 0 0 1  L 2c 21  L11   ( L1  L 2c 2)1 

50
51
Kinematics Relations - Joint & Cartesian Spaces

 A robot is often used to manipulate object attached to its tip (end

effector).
 The location of the robot tip may be specified using one of the
following descriptions:

 Joint Space
 1  {N}
 
  2

 
 N 

 Cartesian Space

 R
0 0
PN   
 0 PN 
X  0 
T 
 
0 N
N 
 0 1   rN 
52
Kinematics Relations – Forward & Inverse

 The robot kinematic equations relate the two description of the

robot tip location

X  FK ( )
 1 
 
  2  
 0 PN 
X  0 

 
 
 rN 
 N 

  IK ( X )
Tip Location in Tip Location in
Joint Space Cartesian Space

53
Kinematics Relations – Forward & Inverse

 vx 
 1  v 
   y
d 
  [ ]   2  d  
 v  v 
X  [ X ]   N    z 
dt 
 
dt  
  N   x 
 N   y 
 
 z 
Tip Velocity in Tip velocity in
Joint Space Cartesian Space

54
 Jacobian Matrix - Introduction

 The Jacobian is a multi dimensional form of the derivative.

 Suppose that for example we have 6 functions, each of which is a
function of 6 independent variables

y1  f1 ( x1 , x2 , x3 , x4 , x5 , x6 )
y2  f 2 ( x1 , x2 , x3 , x4 , x5 , x6 )

y6  f 6 ( x1 , x2 , x3 , x4 , x5 , x6 )

 We may also use a vector notation to write these equations as

Y  F(X )

55
 Jacobian Matrix - Introduction
 If we wish to calculate the differential of yi as a function of the
differential xi we use the chain rule to get

f1 f f
y1  x1  1 x2    1 x6
x1 x2 x6
f 2 f f
y 2  x1  2 x2    2 x6
x1 x2 x6

f 6 f f
y 6  x1  6 x2    6 x6
x1 x2 x6

 Which again might be written more simply using a vector notation as

F
Y  X
X
56
 Jacobian Matrix - Introduction
 The 6x6 matrix of partial derivative is defined as the Jacobian
matrix
F
Y  X  J ( X )X
X

 By dividing both sides by the differential time element, we can think

of the Jacobian as mapping velocities in X to those in Y

Y  J ( X ) X

 Note that the Jacobian is time varying linear transformation

57
 Jacobian Matrix - Introduction

 In the field of robotics the

Jacobian matrix describe the
relationship between the joint
angle rates (  N ) and the
translation and rotation velocities
of the end effector ( x  ). This
relationship is given by:

x  J  

  J  1 x

58
 Jacobian Matrix - Introduction
 This expression can be expanded to:

 x     1 
 y     
     2 
 z   J    
   
 x    
 y    
     
 z     N 
6x1 6xN Nx1

 Where:
 x is a 6x1 vector of the end effector linear and angular velocities
 J   is a 6xN Jacobian matrix
  N is a Nx1 vector of the manipulator joint velocities
 N is the number of joints
59
 Jacobian Matrix - Introduction
 In addition to the velocity
relationship, we are also interested
in developing a relationship
between the robot joint torques (  )
F
and the forces and moments ( F )
at the robot end effector (Static
Conditions). This relationship is
given by:

  J  T F

60
 Jacobian Matrix - Introduction
 This expression can be expanded to:

 1  
T
  Fx 
    F 
 2    y
   J     Fz 
    
    M x 
    M y 
     
 N    M z 
6x1 6xN Nx1

 Where:
  is a 6x1 vector of the robot joint torques
J   is a 6xN Transposed Jacobian matrix
T

 F is a Nx1 vector of the forces and moments at the robot end
effector
61  N is the number of joints