Introduction to Robotics

1 1/12/2018
General Course Information
 The course introduces the basic concepts of robotic
manipulators and autonomous systems. After a review of
some fundamental mathematics the course examines the
mechanics and dynamics of robot arms, mobile robots, their
sensors and algorithms for controlling them.
 two robotic arms
 everything in Matlab (and some Java)

2 1/12/2018
A150 Robotic Arm

link 3
link 2

3 Symbolic Representation of Manipulators 1/12/2018

Kinematics
 the study of motion that ignores the forces that cause the
motion
 “geometry of motion”
 interested in position, velocity, acceleration, etc. of the various
links of the manipulator
 e.g., where is the gripper relative to the base of the manipulator?
what direction is it pointing in?
 described using rigid transformations of the links

4 1/12/2018
Kinematics
 forward kinematics:
 given the link lengths and joint angles compute the position and
orientation of the gripper relative to the base
 for a serial manipulator there is only one solution
 inverse kinematics:
 given the position (and possibly the orientation) of the gripper and
the dimensions of the links, what are the joint variables?
 for a serial manipulator there is often more than one mathematical solution

5 1/12/2018
Wheeled Mobile Robots
 robot can have one or more wheels that can provide
 steering (directional control)
 power (exert a force against the ground)
 an ideal wheel is
 perfectly round (perimeter 2πr)
 moves in the direction perpendicular to its axis

6 1/12/2018
Wheel

7 1/12/2018
Deviations from Ideal

8 1/12/2018
Differential Drive
 two independently driven wheels mounted on a common axis

9 1/12/2018
Forward Kinematics
 for a robot starting with pose [0 0 0]T moving with velocity
V(t) in a direction θ(t) :

t
x(t )   V (t ) cos( (t )) dt
0
t
y (t )   V (t ) sin(  (t )) dt
0
t
 (t )    (t ) dt
0

10 1/12/2018
Sensitivity to Wheel Velocity

vr (t )  1  N (0,  2 )
v (t )  1  N (0,  2 )
 (0)  0 σ = 0.05

t  010
  0.2 σ = 0.01

11 1/12/2018
Localization using Landmarks: RoboSoccer

12 1/12/2018
Maps

* goal

start *

13 1/12/2018
Path Finding

goal

start

14 1/12/2018
Localization

goal

start

15 1/12/2018
EKF SLAM Application

www.probabilistic-robotics.org [MIT B21, courtesy by John Leonard]

16
EKF SLAM Application

raw odometry estimated trajectory

www.probabilistic-robotics.org [courtesy by John Leonard]

17
 Day 02

Introduction to manipulator kinematics

18 1/12/2018
Robotic Manipulators
 a robotic manipulator is a kinematic chain
 i.e. an assembly of pairs of rigid bodies that can move respect to one
another via a mechanical constraint
 the rigid bodies are called links
 the mechanical constraints are called joints

19 Symbolic Representation of Manipulators 1/12/2018

A150 Robotic Arm

link 3
link 2

20 Symbolic Representation of Manipulators 1/12/2018

Joints
 most manipulator joints are one of two types
1. revolute (or rotary)
 like a hinge
 allows relative rotation about a fixed axis between two links
 axis of rotation is the z axis by convention
2. prismatic (or linear)
 like a piston
 allows relative translation along a fixed axis between two links
 axis of translation is the z axis by convention
 our convention: joint i connects link i – 1 to link i
 when joint i is actuated, link i moves

21 Symbolic Representation of Manipulators 1/12/2018

Joint Variables
 revolute and prismatic joints are one degree of freedom
(DOF) joints; thus, they can be described using a single
numeric value called a joint variable
 qi : joint variable for joint i
1. revolute
 qi = i : angle of rotation of link i relative to link i – 1
2. prismatic
 qi = di : displacement of link i relative to link i – 1

22 Symbolic Representation of Manipulators 1/12/2018

Revolute Joint Variable
 revolute
 qi = i : angle of rotation of link i relative to link i – 1

link i

i
link i – 1

23 Symbolic Representation of Manipulators 1/12/2018

Prismatic Joint Variable
 prismatic
 qi = di : displacement of link i relative to link i – 1

link i – 1 link i

di

24 Symbolic Representation of Manipulators 1/12/2018

Common Manipulator Arrangments
 most industrial manipulators have six or fewer joints
 the first three joints are the arm
 the remaining joints are the wrist
 it is common to describe such manipulators using the joints of
the arm
 R: revolute joint
 P: prismatic joint

25 Common Manipulator Arrangements 1/12/2018

Articulated Manipulator
 RRR (first three joints are all revolute)
 joint axes
 z0 : waist
 z1 : shoulder (perpendicular to z0) z0
 z2 : elbow (parallel to z1) z1 z2

2 3
shoulder
forearm
elbow
1

waist

26 Common Manipulator Arrangements 1/12/2018

Spherical Manipulator
 RRP
 Stanford arm
 http://infolab.stanford.edu/pub/voy/museum/pictures/display/robots/IMG_2404ArmFrontPeekingOut.JPG

z0
z1
d3

2
shoulder z2

1

waist

27 Common Manipulator Arrangements 1/12/2018

SCARA Manipulator
 RRP
 Selective Compliant Articulated Robot for Assembly
 http://www.robots.epson.com/products/g-series.htm

z1 z2

z0
2

d3

1

28 Common Manipulator Arrangements 1/12/2018

Forward Kinematics
 given the joint variables and dimensions of the links what is
the position and orientation of the end effector?

a2

2

a1

1

29 Forward Kinematics 1/12/2018

Forward Kinematics
 choose the base coordinate frame of the robot
 we want (x, y) to be expressed in this frame

(x, y) ?

a2

2
y0

a1

1
x0

30 Forward Kinematics 1/12/2018

Forward Kinematics
 notice that link 1 moves in a circle centered on the base frame
origin

(x, y) ?

a2

2
y0

a1

1 ( a1 cos 1 , a1 sin 1 )
x0

31 Forward Kinematics 1/12/2018

Forward Kinematics
 choose a coordinate frame with origin located on joint 2 with
the same orientation as the base frame

(x, y) ?

y1
a2

2
y0
1
a1 x1

1 ( a1 cos 1 , a1 sin 1 )
x0

32 Forward Kinematics 1/12/2018

Forward Kinematics
 notice that link 2 moves in a circle centered on frame 1

(x, y) ?

y1
a2
( a2 cos (1 + 2),
2 a2 sin (1 + 2) )
y0
1
a1 x1

1
( a1 cos 1 , a1 sin 1 )
x0

33 Forward Kinematics 1/12/2018

Forward Kinematics
 because the base frame and frame 1 have the same
orientation, we can sum the coordinates to find the position
of the end effector in the base frame (a1 cos 1 + a2 cos (1 + 2),
a1 sin 1 + a2 sin (1 + 2) )

y1
a2
( a2 cos (1 + 2),
2 a2 sin (1 + 2) )
y0
1
a1 x1

1
( a1 cos 1 , a1 sin 1 )
x0

34 Forward Kinematics 1/12/2018

Forward Kinematics
 we also want the orientation of frame 2 with respect to the
base frame
 x2 and y2 expressed in terms y2 x2
of x0 and y0

a2

2
y0
1
a1

1
x0

35 Forward Kinematics 1/12/2018

Forward Kinematics
 without proof I claim:
x2 = (cos (1 + 2),
y2
sin (1 + 2) ) x2

y2 = (-sin (1 + 2),

cos (1 + 2) )
a2

2
y0
1
a1

1
x0

36 Forward Kinematics 1/12/2018

Inverse Kinematics
 given the position (and possibly
the orientation) of the end
effector, and the dimensions y2 x2
of the links, what are the joint (x, y)
variables?
a2

2 ?
y0

a1

1 ?
x0

37 Inverse Kinematics 1/12/2018

Inverse Kinematics
 harder than forward kinematics because there is often more
than one possible solution

(x, y)

a2

y0

a1

x0

38 Inverse Kinematics 1/12/2018

Inverse Kinematics
law of cosines
b 2  a12  a22  2a1a2 cos(   2 )  x 2  y 2

(x, y)

a2
b
2 ?
y0

a1

x0

39 Inverse Kinematics 1/12/2018

Inverse Kinematics

x 2  y 2  a12  a22
 cos(   2 ) 
2a1a2
and we have the trigonometric identity

 cos(  2 )  cos(2 )
therefore,

x 2  y 2  a12  a22
cos  2   C2
2a1a2

We could take the inverse cosine, but this gives only one of the two solutions.

40 Inverse Kinematics 1/12/2018

Inverse Kinematics

Instead, use the two trigonometric identities:

sin 
sin   cos  2  1
2 2
tan  
cos 

to obtain

 1  C 2
 2  tan 1 2

C2
which yields both solutions for 2 . In many programming languages you would use the
four quadrant inverse tangent function atan2

c2 = (x*x + y*y – a1*a1 – a2*a2) / (2*a1*a2);

s2 = sqrt(1 – c2*c2);
theta21 = atan2(s2, c2);
theta22 = atan2(-s2, c2);

41 Inverse Kinematics 1/12/2018

Inverse Kinematics
 Exercise for the student: show that

 y 1  a2 sin  2 
1  tan    tan 
1

 x  a1  a2 cos  2 

42 Inverse Kinematics 1/12/2018

 Day 03

Spatial Descriptions

43 1/12/2018
Points and Vectors
 point : a location in space
 vector : magnitude (length) and direction between two points

v
q

44 1/12/2018
Coordinate Frames
 choosing a frame (a point and two perpendicular vectors of
unit length) allows us to assign coordinates

0
p

4
p  
0
ŷ0 q0 2.5
2
q  
0

x̂0 1.5
0 o0  2
v  p q   
0 0 0

1 
45 1/12/2018
Coordinate Frames
 the coordinates change depending on the choice of frame

1
p

1 o1
ŷ1  0.5
p 
1

q1  4 
0.5
x̂1 q  
1

2
 1
v  p q   
1 1 1

2
46 1/12/2018
Dot Product
 the dot product of two vectors
 u1   v1 
u  v 
u   2 v   2 u  v  u1v1  u2v2    un vn  u T v
 
    u  v  u v cos 
un  vn 


v
u cos

47 1/12/2018
Translation

ŷ0 ŷ1

x̂0 x̂1
o0 o1
0 
1
 suppose we are given o1 expressed in {0}

 3
o  
0
1
0 

48 1/12/2018
Translation 1

ŷ0 ŷ1

x̂0 x̂1
o0 o1
0 
1
 the location of {1} expressed in {0}

3 0 3

d  o o       
0
1
0
1
0
0
 0  0  0 

49 1/12/2018
Translation 1
1. the translation vector d ij can be interpreted as the location
of frame {j} expressed in frame {i}

50 1/12/2018
Translation 2
a point expressed
in frame {1}

ŷ0 ŷ1 1

p  
1

1
x̂0 x̂1
o0 o1
0 
1
 p1 expressed in {0}

3 1 4

p d  p     
0 0
1
1

0 1 1

51 1/12/2018
Translation 2
2. the translation vector d ij can be interpreted as a coordinate
transformation of a point from frame {j} to frame {i}

52 1/12/2018
Translation 3

 1 ŷ0  2
p  
0
q  
0

1 1 
x̂0
o0
0
 q0 expressed in {0}

3  1 2
q d  p     
0 0

 0   1  1 

53 1/12/2018
Translation 3
3. the translation vector d can be interpreted as an operator
that takes a point and moves it to a new point in the same
frame

54 1/12/2018
Rotation
 suppose that frame {1} is rotated relative to frame {0}

ŷ0
ŷ1
x̂1
sin 

o0  o1 x̂0

cos

55 1/12/2018
Rotation 1
 the orientation of frame {1} expressed in {0}

ŷ0
ŷ1
x̂1


o0  o1 x̂0

 x1  x0 y1  x0 
R 
0

 x1  y0 y1  y0 
1

56 1/12/2018
Rotation 1
1. the rotation matrix R ij can be interpreted as the orientation
of frame {j} expressed in frame {i}

57 1/12/2018
Rotation 2
 p1 expressed in {0}
1
p  
1
ŷ0
1
ŷ1
x̂1


o0  o1 x̂0

cos   sin   1

p R p 
0 0 1

 sin  cos   1

1

58 1/12/2018
Rotation 2
2. the rotation matrix R ij can be interpreted as a coordinate
transformation of a point from frame {j} to frame {i}

59 1/12/2018
Rotation 3
 q0 expressed in {0}

q0
ŷ0
0
p

o0 x̂0

cos   sin   1

q R p 
0 0

 sin  cos   1

60 1/12/2018
Rotation 3
3. the rotation matrix R can be interpreted as an operator
that takes a point and moves it to a new point in the same
frame

61 1/12/2018
Properties of Rotation Matrices
 RT = R-1
 the columns of R are mutually orthogonal
 each column of R is a unit vector
 det R = 1 (the determinant is equal to 1)

62 1/12/2018
Rotation and Translation

ŷ1
ŷ0 x̂1
o1
x̂0 
1
o0
0

63 1/12/2018
Rotations in 3D

 x1  x0 y1  x0 z1  x0 
R10   x1  y0 y1  y0 z1  y0 
 x1  z0 y1  z0 z1  z0 

64 1/12/2018
 Day 04

Rotations

65 1/12/2018
Properties of Rotation Matrices
 RT = R-1
 the columns of R are mutually orthogonal
 each column of R is a unit vector
 det R = 1 (the determinant is equal to 1)

66 1/12/2018
Rotations in 3D

ẑ0
ŷ1
ẑ1
x̂1

ŷ0
o0  o1

 x1  x0 y1  x0 z1  x0 
x̂0
R10   x1  y0 y1  y0 z1  y0 
 x1  z0 y1  z0 z1  z0 

67 1/12/2018
Rotation About z-axis

zˆ0  zˆ1
+'ve rotation

ŷ1
ŷ0


x̂0 x̂1

68 1/12/2018
Rotation About x-axis

ẑ0
ẑ1

ŷ1

 ŷ0
+'ve rotation

x̂0 x̂1

69 1/12/2018
Rotation About y-axis

ẑ0

ẑ1

yˆ 0  yˆ1

+'ve rotation

x̂0 x̂1

70 1/12/2018
Relative Orientation Example

ẑ0
x̂1
45

x̂0

ŷ1
yˆ 0  zˆ1

71 1/12/2018
Successive Rotations in Moving Frames
1. Suppose you perform a rotation in frame {0} to obtain {1}.
2. Then you perform a rotation in frame {1} to obtain {2}.

What is the orientation of {2} relative to {0}?

ẑ1 ẑ0 zˆ1  zˆ2 ẑ2 ẑ0


x̂0 x̂0
 ŷ2 ŷ2
x̂1
x̂1
yˆ 0  yˆ1 x̂2 ŷ1 x̂2 ŷ0
R10  Ry , R21  Rz , R20  ?
72 1/12/2018
Successive Rotations in a Fixed Frame
1. Suppose you perform a rotation in frame {0} to obtain {1}.
2. Then you rotate {1} in frame {0} to obtain {2}.

What is the orientation of {2} relative to {0}?


ẑ1 ẑ0 ẑ1 ẑ0 ẑ2 ẑ0
ŷ1 ŷ2
x̂0 x̂0 x̂0

x̂1
yˆ 0  yˆ1 x̂1 ŷ0 x̂2 ŷ0
R10  Ry , R  Rz , R20  ?
73 1/12/2018
Composition of Rotations
1. Given a fixed frame {0} and a current frame {1} and R10
 if {2} is obtained by a rotation R in the current frame {1} then use
postmulitplication to obtain:

RR 1
2 and R  R R
0
2
0
1
1
2

2. Given a fixed frame {0} and a frame {1} and

 if {2} is obtained by a rotation R in the fixed frame {0} then use
premultiplication to obtain:

R  RR
0
2
0
1

74 1/12/2018
Rotation About a Unit Axis

ẑ0
kz
k̂

ky c  cos 
ŷ0
kx s  sin 
v  1  cos 
x̂0
 k x2 v  c k x k y v  k z s k x k z v  k y s 
 
Rk ,  k x k y v  k z s k y2 v  c k y k z v  k x s 
k x k z v  k y s k y k z v  k x s k z2 v  c 

75 1/12/2018
 Day 05

Rigid Body Transformations

76 1/12/2018
Homogeneous Representation
 translation represented by a vector d
 vector addition
 rotation represented by a matrix R
 matrix-matrix and matrix-vector multiplication
 convenient to have a uniform representation of translation
and rotation
 obviously vector addition will not work for rotation
 can we use matrix multiplication to represent translation?

77 1/12/2018
Homogeneous Representation
 consider moving a point p by a translation vector d

 px  d x   px  d x 
     
p  d   p y   d y    p y  d y 
 pz   d z   pz  d z 

   px   px  d x 
 ?  p   p d 
   y  y y
   pz   pz  d z 

not possible as matrix-vector multiplication always leaves the origin unchanged

78 1/12/2018
Homogeneous Representation
 consider an augmented vector ph and an augmented matrix D
 px  1 0 0 dx 
p  0
 y 1 0 d y 
ph  D
 pz  0 0 1 dz 
   
1 0 0 0 1

1 0 0 d x   px   px  d x 
0 1 0 d y   p y   p y  d y 
Dp h   
0 0 1 d z   pz   pz  d z 
    
0 0 0 1  1   1 
79 1/12/2018
Homogeneous Representation
 the augmented form of a rotation matrix R3x3
  0
  0
  R3x3  
R
   0
 0 0 0 1


  0  p x    

R3x3  0  p    R p  
Rp h       y     3x3  
  0  p z     
 0 0 0   
 1  1   1 


80 1/12/2018
Rigid Body Transformations in 3D

{1}
d 10
{0} d 12

{2}

81 1/12/2018
Rigid Body Transformations in 3D
 suppose {1} is a rotated and translated relative to {0}
 what is the pose (the orientation and position) of {1}
expressed in {0} ?
T 10  ?

{1}
d
{0}

82 1/12/2018
Rigid Body Transformations in 3D
 suppose we use the moving frame interpretation (postmultiply
transformation matrices)
1. translate in {0} to get {0’} D00'
2. and then rotate in {0’} to get {1} D00' R10'

{0’}
d
{0}
{0’}
d {1}

Step 1 {0}

Step 2
83 1/12/2018
Rigid Body Transformations in 3D
 suppose we use the fixed frame interpretation (premultiply
transformation matrices)
1. rotate in {0} to get {0’} R
2. and then translate in {0} in to get {1} DR

{0}

{0’} d {1}

Step 1 {0}

Step 2 {0’}
84 1/12/2018
Rigid Body Transformations in 3D
 both interpretations yield the same transformation

T10  D R
1 0 0      0
      0
 0 1 0 d     R3x3  

0 0 1       0
0 0 0 1   0 0 0 1

   
  d  
 R  
  3x3 
     
 0 0 0 1 

85 1/12/2018
Homogeneous Representation
 every rigid-body transformation can be represented as a
rotation followed by a translation in the same frame
 as a 4x4 matrix

 R d
T  
0 0 0 1 

where R is a 3x3 rotation matrix and d is a 3x1 translation vector

86 1/12/2018
Homogeneous Representation
 in some frame i
 points

 p i

P  
i

1

 vectors

v i

V  
i

0

87 1/12/2018
Inverse Transformation
 the inverse of a transformation undoes the original
transformation
 if
 R d
T  
0 0 0 1 
 then

1  R T
 R T
d
T  
0 0 0 1 

88 1/12/2018
 Day 06

Forward Kinematics

89 1/12/2018
Transform Equations

T 12 {2}
{1}

T 10
4
T 2

{0}
0
T 3
T 34
{4}
{3}

90 1/12/2018
Transform Equations
 give expressions for:
0
T 2

3
T 4

91 1/12/2018
Transform Equations

{1}

T 10

{0}
2
T 3
{3}
0
T 2

{2}

92 1/12/2018
Transform Equations
 how can you find

T 10
0
T 2

2
T 3

T 13

93 1/12/2018
Links and Joints
joint n
joint 3 joint 4 joint n-1

link 3 .................
joint 1

link 2
link n-1

link 1 link n

joint 2
link 0
i revolute
qi  
 n joints, n + 1 links di prismatic
 link 0 is fixed (the base)
 joint i connects link i – 1 to link i
 link i moves when joint i is actuated

94 1/12/2018
Forward Kinematics
 given the joint variables and dimensions of the links what is
the position and orientation of the end effector?

p0 ?

a2

2
y0

a1

1
x0

95 Forward Kinematics 1/12/2018

Forward Kinematics
 because the base frame and frame 1 have the same
orientation, we can sum the coordinates to find the position
of the end effector in the base frame (a1 cos 1 + a2 cos (1 + 2),
a1 sin 1 + a2 sin (1 + 2) )

y1
a2
( a2 cos (1 + 2),
2 a2 sin (1 + 2) )
y0
1
a1 x1

1
( a1 cos 1 , a1 sin 1 )
x0

96 Forward Kinematics 1/12/2018

Forward Kinematics
 from Day 02
p0 = (a1 cos 1 + a2 cos (1 + 2),
a1 sin 1 + a2 sin (1 + 2) ) y2 x2

x2 = (cos (1 + 2),

sin (1 + 2) )
a2
y2 = (-sin (1 + 2),
cos (1 + 2) )
2
y0
1
a1

1
x0

97 Forward Kinematics 1/12/2018

Frames

y2 x2

y1
a2

2
y0 x1

a1

1
x0

98 Forward Kinematics 1/12/2018

Forward Kinematics
 using transformation matrices

T  Rz ,1 Dx, a1
0
1

T  Rz ,2 Dx, a2
1
2

T T T
0
2
0
1
1
2

99 1/12/2018
 Day 07

Denavit-Hartenberg

100 1/12/2018
Links and Joints
joint n
joint 3 joint 4 joint n-1

link 3 .................
joint 1

link 2
link n-1

link 1 link n

joint 2
link 0
i revolute
qi  
 n joints, n + 1 links di prismatic
 link 0 is fixed (the base)
 joint i connects link i – 1 to link i
 link i moves when joint i is actuated

101 1/12/2018
Forward Kinematics
 attach a frame {i} to link i
 all points on link i are constant when expressed in {i}
 if joint i is actuated then frame {i} moves relative to frame {i - 1}
 motion is described by the rigid transformation
i 1
T i

 the state of joint i is a function of its joint variable qi (i.e., is a function of qi)

T ii 1 T ii 1(qi )
 this makes it easy to find the last frame with respect to the
base frame
n 1
T 0n  T 10 T 12 T 32  T n

102 1/12/2018
Forward Kinematics
 more generally

T ii 1 T ij12  T j 1
j if i  j
i 
T j  I if i  j

 T j 
i 1
if i  j

 the forward kinematics problem has been reduced to matrix

multiplication

103 1/12/2018
Forward Kinematics
 Denavit J and Hartenberg RS, “A kinematic notation for lower-
pair mechanisms based on matrices.” Trans ASME J. Appl. Mech,
23:215–221, 1955
 described a convention for standardizing the attachment of frames
on links of a serial linkage
 common convention for attaching reference frames on links of
a serial manipulator and computing the transformations
between frames

104 1/12/2018
Denavit-Hartenberg

105 1/12/2018
Denavit-Hartenberg
 notice the form of the rotation component

ci  si c i si s i 

 
 si ci c i  ci s i 
0 s c 
  i  i 

 this does not look like it can represent arbitrary rotations

 can the DH convention actually describe every physically
possible link configuration?

106 1/12/2018
Denavit-Hartenberg
 yes, but we must choose the orientation and position of the
frames in a certain way

 (DH1) xˆ1  zˆ0

 (DH2) xˆ1 intersects zˆ0

 claim: if DH1 and DH2 are true then there exists unique
numbers

a, d ,  ,  such that T10  R z , Dz ,d Dx,a R x,

107 1/12/2018
Denavit-Hartenberg

T ii 1 Rz ,i Tz ,di Tx,ai Rx,i

ci  si c i si s i ai ci 
s c c  c s a s 
  i i  i i  i i i 

0 s i c i di 
 
0 0 0 1 
ai link length
i link twist
di link offset
i joint angle
108 1/12/2018
Denavit-Hartenberg
 proof: on blackboard in class

109 1/12/2018
 Day 08

Denavit-Hartenberg

110 1/12/2018
Denavit-Hartenberg Forward Kinematics
 RPP cylindrical manipulator
 http://strobotics.com/cylindrical-format-robot.htm

111 1/12/2018
Denavit-Hartenberg Forward Kinematics
How do we place the frames?

112 1/12/2018
Step 1: Choose the z-axis for each frame
 recall the DH transformation matrix
i 1
T  Rz ,i Tz ,di Tx,ai Rx,i
i

ci  si c i si s i ai ci 

s c c  c s a s 
  i i  i i  i i i 

0 s i c i di 
 
0 0 0 1 

xˆii 1 yˆ ii 1 zˆii 1

113 1/12/2018
Step 1: Choose the z-axis for each frame
 ẑi  axis of actuation for joint i+1

link i link i+1 link i link i+1

ẑi

joint i+1 joint i+1

ẑi

114 1/12/2018
Step 1: Choose the z-axis for each frame

 Warning: the picture is deceiving. We do not yet know the origin of the
frames; all we know at this point is that each zi points along a joint axis

115 1/12/2018
Step 2: Establish frame {0}
 place the origin o0 anywhere on z0
 often the choice of location is obvious
 choose x0 and y0 so that {0} is right-handed
 often the choice of directions is obvious

116 1/12/2018
Step 2: Establish frame {0}

117 1/12/2018
Step 3: Iteratively construct {1}, {2}, ... {n-1}
 using frame {i-1} construct frame {i}
 DH1: xi is perpendicular to zi-1
 DH2: xi intersects zi-1
 3 cases to consider depending on the relationship between zi-1
and zi

118 1/12/2018
Step 3: Iteratively construct {1}, {2}, ... {n-1}
 Case 1
 zi-1 and zi are not coplanar (skew)

zˆi 1
shortest line between ẑi (out of page)
zˆi 1 and ẑi
x̂i
ai

oi point of intersection

 i angle from zi-1 to zi measured about xi

119 1/12/2018
Step 3: Iteratively construct {1}, {2}, ... {n-1}
 Case 2
 zi-1 and zi are parallel ( i = 0 )

zˆi 1 ẑi

oi 1 x̂i
ai
oi point of intersection

 notice that this choice results in di = 0

120 1/12/2018
Step 3: Iteratively construct {1}, {2}, ... {n-1}
 Case 3
 zi-1 and zi intersect ( ai = 0 )

zˆi 1

oi point of intersection
x̂i

(out of page) ẑi

121 1/12/2018
Step 3: Iteratively construct {1}, {2}, ... {n-1}

122 1/12/2018
Step 3: Iteratively construct {1}, {2}, ... {n-1}

123 1/12/2018
Step 4: Place the end effector frame

“sliding”

“normal”

“approach”

124 1/12/2018
Step 4: Place the end effector frame

125 1/12/2018
Step 5: Find the DH parameters
 ai : distance between zi-1 and zi measured along xi
 i : angle from zi-1 and zi measured about xi
 di : distance between oi-1 to the intersection of xi and zi-1
measured along zi-1
 i : angle from xi-1 and xi measured about zi-1

126 1/12/2018
Step 5: Find the DH parameters

Link ai i di i
1 0 0 d1 1*
2 0 -90 d2* 0
3 0 0 d3* 0
* joint variable

127 1/12/2018
 Day 09

More Denavit-Hartenberg Examples

128 1/12/2018
Step 5: Find the DH parameters

Link ai i di i
1 0 0 d1 1*
2 0 -90 d2* 0
3 0 0 d3* 0
* joint variable

129 1/12/2018
Step 6: Compute the transformation
 once the DH parameters are known, it is easy to construct
the overall transformation
Link ai i di i
1 0 0 d1 1*
2 0 -90 d2* 0
3 0 0 d3* 0
* joint variable

c1  s1 0 0
s c 0 0 
T 10  Rz ,1Tz ,d1Tx ,a1 Rx ,1  1 1
0 0 1 d1 
 
0 0 0 1

130 1/12/2018
Step 6: Compute the transformation

Link ai i di i
1 0 0 d1 1*
2 0 -90 d2* 0
3 0 0 d3* 0
* joint variable

1 0 0 0
0 0 1 0 
T 12  Rz , 2 Tz ,d 2 Tx,a2 Rx, 2 
0  1 0 d2 
 
0 0 0 1

131 1/12/2018
Step 6: Compute the transformation

Link ai i di i
1 0 0 d1 1*
2 0 -90 d2* 0
3 0 0 d3* 0
* joint variable

1 0 0 0
0 1 0 0 
T 32  Rz ,3 Tz ,d3 Tx ,a3 Rx , 3 
0 0 1 d3 
 
0 0 0 1

132 1/12/2018
Step 6: Compute the transformation

c1 0  s1  s1d3 
s 0 c c d 
T 30  T 10T 12T 32   1 1 1 3 

 0  1 0 d1  d 2 
 
0 0 0 1 

133 1/12/2018
Spherical Wrist

134 1/12/2018
Spherical Wrist

135 1/12/2018
Spherical Wrist: Step 1

136 1/12/2018
Spherical Wrist: Step 2

137 1/12/2018
Spherical Wrist: Step 2

138 1/12/2018
Spherical Wrist: Step 4

139 1/12/2018
Step 5: DH Parameters

Link ai i di i
4 0 -90 0 4*
5 0 90 0 5*
6 0 0 d6 6*
* joint variable

140 1/12/2018
Step 6: Compute the transformation

c4c5c6  s4 s6  c4c5 s6  s4c6 c4 s5 c4 s5 d 6 

s c c  c s  s c s  c c s s s s d 
T 36  T 34T 54T 56   4 5 6 4 6 4 5 6 4 6 4 5 4 5 6

  s5c6 s5 s6 c5 c5 d 6 
 
 0 0 0 1 

141 1/12/2018
RPP + Spherical Wrist

142 1/12/2018
RPP + Spherical Wrist

 r11 r12 r13 d x 

r r r d 
T 06  T 30T 36   21 22 23 y 
r31 r32 r33 d z 
 
0 0 0 1

r11  c1c4c5c6  c1s4 s6  s1s5c6


d z   s4 s5 d 6  d1  d 2

143 1/12/2018
Stanford Manipulator + Spherical Wrist

Link ai i di i
1 0 -90 0 1*
2 0 90 d2 2*
3 0 0 d3* 0
4 0 -90 0 4*
5 0 90 0 5*
6 0 0 d6 6*
* joint variable

144 1/12/2018
SCARA + 1DOF Wrist

Link ai i di i
1 a1 0 d1 1*
2 a2 180 0 2*
3 0 0 d3* 0
4 0 0 d4 4* * joint variable

145 1/12/2018