CHAPTER 3


ROBOT CLASSIFICATION

DAE 32503 – ROBOTICS & AUTOMATION SYSTEM


By: Nor Faezah Adan
1 ROBOT GEOMETRY

2 KINEMATICS & PLANNING

3 DYNAMICS &CONTROL
REFERENCES

1. Craig, John J. (2005). Introduction to

robotics : Mechanics and control.
Pearson. Shelf No.: TJ211 .C72 2005
Robot Geometry: Degrees of Freedom (DOF)

For each degree of freedom, a joint

is required.
1 joint = 1 DOF
• The more degrees of freedom, the greater the
complexity of motions encountered.
• For applications that require more flexibility,
additional degrees of freedom are used in the
wrist of the robot.
• Three degrees of freedom located in the wrist
give the end effector all the flexibility.
6
DOF
Robot Geometry: Degrees of Freedom (DOF)

How many
DOF?
Robot Geometry: Degrees of Freedom (DOF)

How many
DOF?
Robot Geometry: Degrees of Freedom (DOF)

The 3 DOF located in the arm of a robotic system:

1. The rotational traverse
The rotational traverse is the movement of the
arm assembly about a rotary axis, such as the
left-and-right swivel of the robot’s arm on a
base.
2. The radial traverse
The radial traverse is the extension and
retraction of the arm or the in-and-out motion
relative to the base.
Robot Geometry: Degrees of Freedom (DOF)

3. The vertical traverse

The vertical traverse provides the up-and-
down motion of the arm of the robotic system.
Robot Geometry: Degrees of Freedom (DOF)
Robot Geometry: Degrees of Freedom (DOF)

The 3 DOF located in the wrist of a robotic

system:
1. Pitch
Bend or up and down movement.
2. Yaw
Right and left movement.
3. Roll
Swivel or rotation of the wrist/hand.
Robot Geometry: Degrees of Freedom (DOF)
Robot Geometry: Robot configurations & Work envelope

Robot configurations
• In general, the fundamental mechanical
configurations of robot manipulators are
categorized as Cartesian, Cylindrical,
Spherical and Articulated / Jointed-arm.
• Cartesian is divided into traverse & gantry
types.
• Articulated is divided into horizontal & vertical
types.
Robot Geometry: Robot configurations & Work envelope

Work envelope / workspace

• The extreme position of the robot axes describe a
boundary for the region in which the robot operates.
This boundary encloses the work envelope.
• The size of a work envelope determines the limits of
reach.
Robot Geometry: Robot configurations & Work envelope

1. Cartesian / Rectangular robot

• It has 3 prismatic joints, whose axes are coincident
with a cartesian coordinate system.
• Most cartesian robots come as Gantries,
distinguished by a frame structure supporting the
linear axes.
• Gantry robots are widely used for:
- Special machining tasks such as water jet or laser
cutting where robot motion cover large surfaces.
- Palletizing
- Warehousing
Robot Geometry: Robot configurations & Work envelope

Gant Traver
ry se
Robot Geometry: Robot configurations & Work envelope

Work envelope: Cartesian /

rectangular robot
• Shaped as a cube or a rectangle.
Robot Geometry: Robot configurations & Work envelope
Robot Geometry: Robot configurations & Work envelope

2. Cylindrical robot
• The manipulator has 2 linear motions and 1 rotary
motion.
• Robot’s manipulator has 1 rotational degree of
freedom and 2 translational (linear) degrees of
freedom.
• A cylindrical-coordinated robot generally results in a
larger work envelope than cartesian-coordinated robot.
• This robot is ideally suitable for pick and place
operation.
• Typical applications are assembly, conveyor pallet
transfer, palletizing etc.
Robot Geometry: Robot configurations & Work envelope

Cylindric
al
Robot Geometry: Robot configurations & Work envelope

Work envelope: Cylindrical robot

It can move it’s

gripper within a
volume described
by a cylinder.
Robot Geometry: Robot configurations
Robot Geometry: Robot configurations & Work envelope

3. Spherical / Polar robot

• The manipulator has 1 linear motion and 2 rotary
motions.
• The first motion corresponds to base rotation. The
second motion corresponds to an elbow rotation.
The third motion corresponds to a radial/in-out/
translation.
• A spherical robot generally results in a larger work
envelope than cylindrical and cartesian robot.
• This robot is ideally suitable for applications where
a small amount of vertical movement is adequate
such as loading & unloading a punch press.
Robot Geometry: Robot configurations

Spherica
l
Robot Geometry: Robot configurations & Work envelope

Work envelope: Spherical / Polar robot

The envelope is shaped like a section of a sphere
with upper and lower limits imposed by the
angular rotations of the arm.
Robot Geometry: Robot configurations & Work envelope
Robot Geometry: Robot configurations & Work envelope

4. Articulated / Jointed-arm robot - Vertical

• The manipulator has 3 rotary motions to reach
any point in space.
• The design is similar to human arm.
• The first rotation is about the base, the second
rotation is about the shoulder in a horizontal axis
and the final motion is rotation about the elbow.
• It can move at high speeds and has a greater
variety of angles to approach a given point and
thus very useful for painting and welding
applications.
Robot Geometry: Robot configurations & Work envelope
Robot Geometry: Robot configurations & Work envelope

Work envelope: Vertical articulated/jointed-arm

robot
The envelope is circular when viewed from the top of
the robot. When looked from the side, the envelope has
a circular outer surface with an inner scalloped surface.
Robot Geometry: Robot configurations & Work envelope
Robot Geometry: Robot configurations & Work envelope

4. Articulated / Jointed-arm robot - Horizontal

• The manipulator has 2 rotary motions and 1
linear (vertical) motion to reach any point in
space.
• Also called SCARA (Selective Compliance
Assembly Robot Arm).
• This robot has 2 horizontally jointed-arm
segments fixed to a rigid vertical member (base)
and one vertical linear motion axis.
• It is extremely useful in assembly operations
where insertions of objects into holes are
required.
Robot Geometry: Robot configurations & Work envelope
Robot Geometry: Robot configurations & Work envelope

Work envelope: Horizontal articulated/jointed-arm robot

Robot Geometry: Robot configurations & Work envelope
Configuration Advantages Disadvantages
Cartesian coordinates Three linear axes Can only reach in front of itself
x, y, z (base travel, reach, and Easy to visualize Requires large floor space for
height) Rigid structure size of work envelope
Easy to program off-line Axes hard to seal
Linear axes make for easy
mechanical stops
Cylindrical coordinates Two linear axes, one rotating Cannot reach above itself
θ, y, z – (base rotation, reach, axis Base rotation axis is less rigid
and height) Can reach all around itself than a linear
Reach and height axes rigid Linear axis is hard to seal
axis Won’t reach around obstacles
Rotation axis easy to seal Horizontal motion is circular

Spherical coordinates (vertical) One linear axis, two rotating Can’t reach around obstacles
θ, y, β (base rotation, elevation axes Generally has short vertical
angle, reach angle) Long horizontal reach reach
Articulated (or jointed-arm) Three rotating axes Two or four ways to reach a
coordinates (vertical) Can reach above or below point
θ, β, α (base rotation, elevation obstacles Most complex manipulator
angle, reach angle) Largest work area for least
floor space
SCARA coordinates (horizontal) One linear axis, two rotating Two ways to reach a point
θ, Φ, z (base rotation, reach axes Difficult to program off-line
angle, height) Height axis is rigid Highly complex arm
Large work area for floor
space
Can reach around obstacles

Robot Geometry: Work envelope

Summary of work
envelope
Robot Geometry: Work envelope

End of Lecture 6
Kinematics & Planning: Transformations

Kinematics is the science of motion

that treats motion without regard to the
forces which causes it.
Coordinate systems - Relative frames
Consider the problem of a robot holding a part for
insertion into several CNC machines for various
operations (drilling/grinding). The robot first grasp the
part in a specified way and inserts it into the first
machine. After the first machining operation, the
robot grasps the part in a different way and inserts it
into the second machine. The problem is, how to
exactly specify exactly those 2 gripping positions?
Kinematics & Planning: Transformations

Example:
The vector defining C
is given in {W} frame.
We may need to
transform this into the
robot base co-
ordinate frame {B}
and/or into the end
effector frame {E}.
Kinematics & Planning: Transformations

Point P is located
in coordinate
frame {A}.

The position vector

representing P:
⎡ Px ⎤
A
P = ⎢⎢ Py ⎥⎥
⎢⎣ Pz ⎥⎦
Kinematics & Planning: Transformations

Mappings: Changing descriptions from frame to frame

Describe frames

1. Pure translation
Transform point C vector: Pure translation from frame
{W} to {B}.
**Frame {B} and {W} have the same orientation.
Kinematics & Planning: Transformations

A B A
P = P + PBORG
Kinematics & Planning: Transformations

2. Pure rotation
Transform point C vector: Pure Wrotation
C B
C of
to .
**Frame {B} and {W} have the same origin
position.

Kinematics & Planning: Transformations

Rotation around
z-axis:
B
C X = L1 − L2
B
CY = L3 + L4
B W
CZ = CZ
Kinematics & Planning: Transformations

B W W
C X = C X cos θ − CY sin θ
B
CY = W C X sin θ + W CY cos θ
B W
CZ = CZ
These equations can be expressed
inB matrix form: w
⎡ C X ⎤ ⎡cos θ − sin θ 0⎤ ⎡ C X ⎤
⎢B ⎥ ⎢ ⎥ ⎢w ⎥
⎢ C Y ⎥ = ⎢ sin θ cos θ 0⎥ ⎢ C Y ⎥
⎢ BC Z ⎥ ⎢⎣ 0 0 w
1 ⎦ ⎣ C Z ⎥⎦
⎥ ⎢
⎣ ⎦
B B W
C = RW × C
Kinematics & Planning: Transformations

T
( R ) × R =I
B
W
B
W
−1 T
B
R =( R ) =( R )
W
W
B
W
B
Kinematics & Planning: Transformations

Summary
Rotation matrices:
⎡cos θ − sin θ 0 ⎤
ROT (z, θ ) = ⎢⎢ sin θ cos θ 0 ⎥⎥
⎢⎣ 0 0 1 ⎥⎦
⎡ cos θ 0 sin θ ⎤
ROT (y, θ ) = ⎢⎢ 0 1 0 ⎥⎥
⎢⎣ − sin θ 0 cos θ ⎥⎦
⎡1 0 0 ⎤
ROT (x, θ ) = ⎢⎢0 cos θ − sin θ ⎥⎥
⎢⎣0 sin θ cos θ ⎥⎦

Kinematics & Planning: Transformations

Exercise
1) Pure
translation
Answer
⎡3⎤ ⎡ 2 ⎤ ⎡5 ⎤
B
C = ⎢⎢5 ⎥⎥ + ⎢⎢ −3⎥⎥ = ⎢⎢ 2 ⎥⎥
⎢⎣6 ⎥⎦ ⎢⎣ 1 ⎥⎦ ⎢⎣7 ⎥⎦

Kinematics & Planning: Transformations

Exercise
1) Pure
rotation at
θ=30
Answer
o

⎡cos 30 − sin 30 0 ⎤ ⎡ 4 ⎤
B
C = ⎢⎢ sin 30 cos 30 0 ⎥⎥ ⎢⎢ 2 ⎥⎥
⎢⎣ 0 0 1 ⎥⎦ ⎢⎣ 2 ⎥⎦

⎡ 2.46 ⎤
B
C = ⎢⎢ 3.73 ⎥⎥
⎢⎣ 2 ⎥⎦

Kinematics & Planning: Transformations

3. Combined Translation & Rotation

Frame {B} is not coincident with frame {A}
but has a general vector offset which is the
A
PBORG that locates {B}’s origin,
vector
. A
RB
Also, {B} is rotated with respect to {A}, as
B
P
described by . A
P
Given , we want to compute .

Kinematics & Planning: Transformations

A
P = ( RB × P )+ PBORG
A B A
Kinematics & Planning: Transformations

A
P = ( RB × P )+ PBORG
A B A

A A A B
⎡ P⎤ ⎡ RB PBORG ⎤ ⎡ P ⎤
⎢ ⎥=⎢ ⎥⎢ ⎥
⎢⎣ 1 ⎥⎦ ⎢⎣0 0 0 1 ⎥⎦ ⎢⎣ 1 ⎥⎦

A A B Homogenous Transformation Matrix

P = TB × P or 4x4 Transformation Matrix
Kinematics & Planning: Transformations

Operators: Translations, Rotations & Transformations

 Translate points, rotate vector or both.

1. Translational operators

A
P2 = A P1 + AQ = TRANS (qx , q y , qz )× A P1
⎡1 0 0 qx ⎤
⎢0 1 0 q y ⎥⎥
TRANS (qx , q y , qz )= ⎢
⎢0 0 1 qz ⎥
⎢ ⎥
⎣0 0 0 1⎦
Kinematics & Planning: Transformations

A
2. Rotational operators P2 = R × A P1
⎡cos θ − sin θ 0 0⎤
⎢ sin θ cos θ 0 0 ⎥⎥
ROT (z, θ ) = ⎢
⎢ 0 0 1 0⎥
⎢ ⎥
⎣ 0 0 0 1⎦
⎡ cos θ 0 sin θ 0⎤
⎢ 0 1 0 0 ⎥⎥
ROT (y, θ ) = ⎢
⎢ − sin θ 0 cos θ 0⎥
⎢ ⎥
⎣ 0 0 0 1⎦
⎡1 0 0 0⎤
⎢0 cos θ − sin θ 0 ⎥⎥
ROT (x, θ ) = ⎢
⎢0 sin θ cos θ 0⎥
⎢ ⎥
⎣0 0 0 1⎦
Kinematics & Planning: Transformations

3. Transformation operators
Operator T rotates and translates AaP1 vector
to compute
A
P2 a new vector .

A A
P2 = T × P1
Kinematics & Planning: Transformations

Combined transformations relative to fixed reference frame

A point P(7,3,2)T is attached to a frame {W} and is subjected to

the transformations described below, all relative to a reference
frame {B}. Find the coordinates of the point relative to the fixed
reference frame.
a) Rotation of 90 degrees about z-axis.
b) Followed by rotation of 90 degrees about the y-axis.
c) Followed by a translation of (4,-3,7) T.
Kinematics & Planning: Transformations

B
P = BTW × W P
B
TW = TRANS (x B , y B , z B ) × ROT(θ1 , y B ) × ROT(θ 2 , z B )
(θ1 = 90o , θ 2 = 90o )

B
P = TRANS (x B , y B , z B ) × ROT(θ1 , y B ) × ROT(θ 2 , z B ) × W P
Kinematics & Planning: Transformations

Combined transformations relative to fixed reference frame

A point P(7,3,2)T is attached to a frame {W} and is subjected to

the transformations described below, all relative to a reference
frame {B}. Find the coordinates of the point relative to the fixed
reference frame.
a) Rotation of 90 degrees about z-axis.
b) Followed by a translation of (4,-3,7) T.
c) Followed by rotation of 90 degrees about the y-axis.
Kinematics & Planning: Transformations

B
P = BTW × W P
B
TW = ROT(θ1 , y B ) × TRANS (x B , y B , z B ) × ROT(θ 2 , z B )
(θ1 = 90o , θ 2 = 90o )
B
P = ROT(θ1 , y B ) × TRANS (x B , y B , z B ) × ROT(θ 2 , z B ) × W P
⎡9⎤
⎢4⎥
ANSWER, B P = ⎢ ⎥
⎢ −1⎥
⎢ ⎥
⎣1⎦
Kinematics & Planning: Transformations

Combined transformations relative to rotating frame

A point P(7,3,2)T is attached to a frame {W} and is subjected to

the transformations described below but all relative to the
current moving frame {W}. Find the coordinate of the point
relative to the reference frame {B} after transformations are
completed.
a) Rotation of 90 degrees about z-axis.
b) Followed by rotation of 90 degrees about the y-axis.
c) Followed by a translation of (4,-3,7) T.
Kinematics & Planning: Transformations

B B W
P = TW × P
B
TW = ROT(θ1 , z w ) × ROT(θ 2 , y w ) × TRANS (xW , yW , zW )
(θ1 = 90o , θ 2 = 90o )

B
P = ROT(θ1 , z w ) × ROT(θ 2 , y w ) × TRANS (xW , yW , zW ) × W P
Kinematics & Planning: Transformations

Combined transformations relative to rotating frame

A point P(7,3,2)T is attached to a frame {W} and is subjected to

the transformations described below but all relative to the
current moving frame {W}. Find the coordinate of the point
relative to the reference frame {B} after transformations are
completed.
a) Rotation of 90 degrees about z-axis.
b) Followed by a translation of (4,-3,7) T.
c) Followed by rotation of 90 degrees about the y-axis.
Kinematics & Planning: Transformations

B B W
P = TW × P
B
TW = ROT(θ1 , z w ) × TRANS (xW , yW , zW ) × ROT(θ 2 , y w )
(θ1 = 90o , θ 2 = 90o )
B
P = ROT(θ1 , z w ) × TRANS (xW , yW , zW ) × ROT(θ 2 , y w ) × W P
⎡0 ⎤
⎢6 ⎥
ANSWER, B P = ⎢ ⎥
⎢0 ⎥
⎢ ⎥
⎣1 ⎦
Kinematics & Planning: Transformations

Compound transformations
Kinematics & Planning: Transformations

B {B}.B C
Frame {C} is relative to frameP = TC × P
Frame {B} is relative to frame
A {A}.A B
P = TB × P
A A B C
A B B A
P = TB × TC × P
TB × TC ≠ TC × TB A A B
TC = TB × TC

A
⎡
TC = ⎢
A
RB × RC B
(R
A
B × PCORG )+ PBORG ⎤
B A

⎥
⎢⎣0 0 0 1 ⎥⎦
Kinematics & Planning: Transformations

Inverting a transformation matrix

Consider frame {B} that is known relative to frame
{A}, i.e we know the value of ATB , sometimes, we
will wish to invert this transform in order to get
description of {A} relative to {B}.
⎡ R P⎤
T =⎢ ⎥
⎣0 0 0 1 ⎦
T T
−1
⎡ R − R × P⎤
T =⎢ ⎥
⎢⎣0 0 0 1 ⎥⎦ End of Lecture 7

−1
T ×T = I
Kinematics & Planning: Forward Kinematics

Forward kinematics
Forward kinematics is used to determine the location of
an end effector with respect to a reference coordinate
frame.

Denavit Hartenberg convention (D-H)

Assumptions:
• Robots may be made of a succession of joint and links.
• Joints may be either prismatic or revolute.
• Joints may be in any order or sequence and may be in any
plane.
• Links may also be of any length including zero, maybe
twisted or bent and may be in any plane.
Kinematics & Planning: Forward Kinematics

Assume we are at local reference frame Xn - Zn, we will do

the following four standard steps to get to the next local
reference frame Xn+1 - Zn+1 .
1. Rotate about Zn an angle of θn+1 (this will make Xn and
Xn+1 parallel to each other).
2. Translate along Zn a distance dn+1 to make Xn and Xn+1
collinear.
3. Translate along Xn+1 a distance of an+1 to bring the origins
of both frames together.
4. Rotate Zn axis about Xn+1 axis an angle of αn+1 to align Zn
axis with axis Zn+1 .
Kinematics & Planning: Forward Kinematics

Positive rotation is clockwise

Kinematics & Planning: Forward Kinematics
Kinematics & Planning: Forward Kinematics

n
Tn +1 = ROT ( z , θ n +1 ) × TRANS(0, 0, d n +1 ) × TRANS(a n +1 , 0, 0) × ROT (x, α n +1 )
⎡Cθ n +1 − Sθ n +1 0 0 ⎤ ⎡1 0 0 0 ⎤ ⎡1 0 0 an +1 ⎤ ⎡1 0 0 0⎤
⎢ Sθ Cθ n +1 0 0 ⎥⎥ ⎢⎢0 1 0 0 ⎥⎥ ⎢⎢0 1 0 0 ⎥⎥ ⎢⎢0 Cα n +1 − Sα n +1 0 ⎥⎥
⎢ n +1 × × ×
⎢ 0 0 1 0 ⎥ ⎢0 0 1 d n +1 ⎥ ⎢0 0 1 0 ⎥ ⎢0 Sα n +1 Cα n +1 0⎥
⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ ⎥
⎣ 0 0 0 1 ⎦ ⎣0 0 0 1 ⎦ ⎣0 0 0 1 ⎦ ⎣0 0 0 1⎦

⎡Cθ n +1 − Sθ n +1Cα n +1 Sθ n +1Sα n +1 an +1Cθ n +1 ⎤

n = 1, 2,3...(joint_number)
⎢ Sθ Cθ n +1Cα n +1 −Cθ n +1Sα n +1 an +1Sθ n +1 ⎥⎥
n
Tn +1 = ⎢ n +1 C = cos
⎢ 0 Sα n +1 Cα n +1 d n +1 ⎥
⎢ ⎥ S = sin
⎣ 0 0 0 1 ⎦
R
TH = RT1 1T2 2T3 ....... n −1Tn = A1 A2 A3 ....... An

0
Example: 3 DOF T3 = 0T1 1T2 2T3 = A1 × A2 × A3
robot 
Kinematics & Planning: Forward Kinematics
Kinematics & Planning: Forward Kinematics
Kinematics & Planning: Inverse Kinematics

Inverse kinematics
Inverse kinematics is the calculation of the joint variables from the end effector position and orientation.

In forward kinematic, given the overall transformation matrix is

In inverse kinematics, we will pre-multiply the RT matrix with individual A -1 in order to calculate the angles, θ .
H n n

R R 1 2 n −1
TH = T1 T2 T3 ....... Tn = A1 A2 A3 ....... An .
Kinematics & Planning: Inverse Kinematics

Example: 3 DOF robot.

R
TH = RT1 1T2 2T3 = A1 × A2 × A3
−1 0
A × T3 = A2 × A3
1

A2 −1 × A1−1 × 0T3 = A3

End of Lecture 8
Kinematics & Planning: Dynamics & Control

Differential motion
Suppose you have a robot welding two pieces
together. For best results, you want the robot
to move at a constant speed. This means that
the differential motions of the hand frame must
be defined to represent a constant speed in a
particular direction. This relates to the
differential motion of the frame. However, the
motion is caused by the robot. Thus, we have
to calculate the speeds of each and every joint
at any instant, such that the total motion

Kinematics & Planning: Dynamics & Control

Differential motions of a frame

1. Differential translations
• A differential translation is a translation of
a frame at differential values.
• They can be represented by TRANS
(dx,dy,dz).
• This means that the frame has moved a
differential amount
⎡1 0 0 along
dx ⎤ the three axes.
⎢0 1 0 dy ⎥
TRANS (dx, dy, dz ) = ⎢ ⎥
⎢0 0 1 dz ⎥
⎢ ⎥
⎣ 0 0 0 1 ⎦

Kinematics & Planning: Dynamics & Control

2. Differential rotations
• A small rotation of the frame.
• Since the rotations are small:
➢ Sin δx = δx (in radians)
➢ Cos δx = 1

Kinematics & Planning: Transformations

Differential rotation
matrices:
⎡ 1 −δ z 0 0 ⎤ ⎡cos θ − sin θ 0 0⎤
⎢δ z 1 0 0 ⎥⎥ ⎢ sin θ cos θ 0 0 ⎥⎥
ROT (z, δ z ) = ⎢ ROT (z, θ ) = ⎢
⎢0 0 1 0⎥ ⎢ 0 0 1 0⎥
⎢ ⎥ ⎢ ⎥
⎣0 0 0 1⎦ ⎣ 0 0 0 1⎦
⎡ 1 0 δ y 0⎤ ⎡ cos θ 0 sin θ 0⎤
⎢ 0 1 0 0 ⎥⎥ ⎢ 0 1 0 0 ⎥⎥
ROT (y, δ y ) = ⎢ ROT (y, θ ) = ⎢
⎢ −δ y 0 1 0⎥ ⎢ − sin θ 0 cos θ 0⎥
⎢ ⎥ ⎢ ⎥
⎣ 0 0 0 1⎦ ⎣ 0 0 0 1⎦
⎡1 0 0 0⎤ ⎡1 0 0 0⎤
⎢0 1 −δ x 0 ⎥⎥ ⎢0 cos θ − sin θ 0 ⎥⎥
ROT (x, δ x ) = ⎢ ROT (x, θ ) = ⎢
⎢0 δ x 1 0⎥ ⎢0 sin θ cos θ 0⎥
⎢ ⎥ ⎢ ⎥
⎣0 0 0 1⎦ ⎣0 0 0 1⎦

Kinematics & Planning: Transformations

Differential rotation
operator = ⎡ 1 −δ z δ y 0⎤
⎢ δz 1 −δ x 0 ⎥⎥
ROT (x, δ x )× ROT (y, δ y )× ROT (z, δ z ) = ⎢
⎢ −δ y δ x 1 0⎥
⎢ ⎥
⎣ 0 0 0 1⎦

Kinematics & Planning: Dynamics & Control

3. Differential transformations
The differential transformation of a frame is
a combination of differential translations &
rotations.
[T + dT ] = [TRANS (dx, dy, dz ) × ROT (k, d θ )][T ]
dT = [TRANS (dx, dy, dz ) × ROT (k, d θ ) ][T ]− [T ]
dT = [TRANS (dx, dy, dz ) × ROT (k, d θ ) − I ][T ]
dT = [Δ ][T ]
⎡ 0 −δ z δ y dx ⎤
⎢ δz 0 −δ x dy ⎥
Δ=⎢ ⎥ Differential operato
⎢ −δ y δ x 0 dz ⎥
⎢ ⎥
⎣ 0 0 0 0⎦

Kinematics & Planning: Dynamics & Control

Interpretation of the differential change

• The differential operator represents the
changes in a frame a result of differential
motions.
• As a result of the changes, the new location
and orientation of the frame is given as:
Tnew = Toriginal + dT

Kinematics & Planning: Dynamics & Control

The differential motions of a robot and its

hand frame
• Previous slides mentioned about changes
made to a frame as a result of differential
motions. This only relates to the frame
changes only, not how they were
accomplished.
• In this section we will relate the changes to
the mechanism that accomplishes the
differential motions i.e we will learn how the
robot’s movements are translated into the

Kinematics & Planning: Dynamics & Control

• Consider the frame T discussed previously

as the hand frame of the welding robot. We
will need to find out how the differential
motions of the joints of the robot would
relate to the differential motions of the hand
frame, dT.
• The
⎡ dx ⎤ ⎡relationship
⎤ ⎡ dθ ⎤between the joint
1
⎢ dy ⎥ ⎢ ⎥ ⎢ dθ ⎥
movements
⎢ ⎥ ⎢ and
⎥⎢ ⎥the hand movement can
2

⎢ dz ⎥ ⎢ ROBOT ⎥ ⎢ dθ ⎥
be
⎢ ⎥=⎢
δ x
⎢ ⎥ ⎢ linked using
JACOBI ⎥⎢
⎥⎢ dθthe
⎥
⎥
3

4
D
robot J
JACOBIAN.D
[ ] [ ][ ]
= θ
⎢δ y ⎥ ⎢ AN ⎥ ⎢ dθ5 ⎥
⎢ ⎥ ⎢ ⎥⎢ ⎥
⎣δ z ⎦ ⎣ dθ
⎦ ⎣⎢ 6 ⎦⎥

Kinematics & Planning: Dynamics & Control

How to relate the JACOBIAN and the

differential operator
Example:
A frame of a robot with 5 DOF, its numerical
Jacobian for this instant, and a set of
differential
⎡1 0 0 motions
5⎤ ⎡ 3are
0 given.
0 0 0 ⎤ Find
⎡ dθthe
1⎤ ⎡new
0.1 ⎤
⎢0 0 −1 3 ⎥ ⎢ −2 0 1 0 0 ⎥ ⎢ dθ ⎥ ⎢ −0.1⎥
location
T6 = ⎢
of the
⎥ hand⎢ after the differential
⎥ ⎢ 2⎥ ⎢ ⎥
⎢0 1 0 2 ⎥
motion. J = ⎢ 0 4 0 0 0⎥ ⎢ dθ3 ⎥ = ⎢0.05⎥
⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ ⎥
⎣0 0 0 1⎦ ⎢ 0 1 0 1 0⎥ ⎢ dθ 4 ⎥ ⎢ 0.1 ⎥
⎢⎣ −1 0 0 0 1 ⎥⎦ ⎢⎣ dθ5 ⎥⎦ ⎢⎣ 0 ⎥⎦

Kinematics & Planning: Dynamics & Control

Solution:
1. Use [D]=[J][Dθ] to find dx,dy,dz,δx and δy.
2. Substitute the values into the differential
operator matrix.
3. Use [dT]=[∆][T] to get the differential
motion of the frame.
4. Use Tnew = Toriginal⎡⎢ 1+ dT −0.1 0 5.1 ⎤
to get the⎥ new
⎢ 0 0 −1 2.85⎥
T6( newframe
location of the ) =
⎢0.1 after
0 0the2.1
differential
⎥
motion. ⎢ ⎥ End of Lecture 9
⎣0 0 0 1 ⎦
5. Final answer 

Thank You
Thank you