Direct Kinematics: Robotics 1

Download as pdf or txt
Download as pdf or txt
You are on page 1of 31

Robotics 1

Direct kinematics
Prof. Alessandro De Luca

Robotics 1

Kinematics of robot manipulators


!

study of ...
geometric and timing aspects of robot motion,
without reference to the causes producing it
robot seen as ...
an (open) kinematic chain of rigid bodies
interconnected by (revolute or prismatic) joints

Robotics 1

Motivations
!

functional aspects
!
!

definition of robot workspace


calibration

operational aspects

task execution
(actuation by motors)

task definition and


performance

two different spaces related by kinematic (and dynamic) maps


!
!
!

Robotics 1

trajectory planning
programming
motion control
3

Kinematics

formulation and parameterizations


r = f(q)
JOINT
space

q = (q1,,qn)
!

q = f -1(r)

r = (r1,,rm)

choice of parameterization q
!
!

INVERSE

TASK
(Cartesian)
space

DIRECT

unambiguous and minimal characterization of robot configuration


n = # degrees of freedom (dof) = # robot joints (rotational or
translational)

choice of parameterization r
!

compact description of position and/or orientation (pose) variables of


interest to the required task
usually, m ! n and m ! 6 (but none of these is strictly necessary)

Robotics 1

Open kinematic chains


q2

q4

q1

q3
e.g., the relative angle
between a link and the
following one
!

r = (r1,,rm)
e.g., it describes the
pose of frame RFE

pointing in space
positioning in the plane

m=3
!
!

Robotics 1

RFE

m=2
!

qn

orientation in space
positioning and orientation in the plane
5

Classification by kinematic type


(first 3 dofs)

SCARA
(RRP)

Cartesian or
gantry
(PPP)

cylindric
(RPP)
articulated or
anthropomorphic
(RRR)
polar or
spherical
(RRP)

Robotics 1

R = 1-dof rotational (revolute) joint


P = 1-dof translational (prismatic) joint

Direct kinematic map


!

the structure of the direct kinematics function


depends from the chosen r

r = fr(q)
!

methods for computing fr(q)


!
!

Robotics 1

geometric/by inspection
systematic: assigning frames attached to the robot
links and using homogeneous transformation matrices

Example: direct kinematics of 2R arm


y

py

l2
l1

"

q2

q1
px

q1
q=
q2

n=2

px
r = py
"

m=3

px = l1 cos q1 + l2 cos(q1+q2)
py = l1 sin q1 + l2 sin(q1+q2)
" = q1+ q2
for more general cases, we need a method!
Robotics 1

Numbering links and joints


joint 2

joint i+1

link i-1

link 1
joint 1
link 0
(base)

joint i-1

revolute
Robotics 1

link i
joint i

joint n
link n
(end effector)

prismatic
9

Spatial relation between joint axes


axis of joint i+1

axis of joint i
90
A

%$

# i$

90
B

common normal
(axis of link i)

a i = displacement AB between joint axes (always well defined)


# i = twist angle between joint axes
projected on a plane % orthogonal to the link axis

Robotics 1

with sign
(pos/neg)!
10

Spatial relation between link axes


axis of joint i

link i-1

link i
D
C
axis of link i-1

axis of link i

&i

'$

d i = displacement CD (a variable if joint i is prismatic)


& i = angle between link axes (a variable if joint i is revolute)
projected on a plane ' orthogonal to the joint axis

Robotics 1

with sign
(pos/neg)!
11

Denavit-Hartenberg (DH) frames


joint axis
i-1

joint axis
i

link i-1

link i

link i-1 is moved


by joint i-1

frame RFi is
attached to link i

zi-1
xi-1

di
Oi-1

axis around which the link rotates


or along which the link slides
Robotics 1

joint axis
i+1

ai

zi
Oi

&i

#i
xi

common normal
to joint axes
i and i+1
12

Denavit-Hartenberg parameters
axis of joint
i

axis of joint
i-1

link i

link i-1
D

zi-1
xi-1

di
Oi-1
!
!
!
!
!
!

ai

axis of joint
i+1

zi
Oi

&i

#i
xi

unit vector zi along axis of joint i+1


unit vector xi along the common normal to joint i and i+1 axes (i ( i+1)
ai = distance DOi positive if oriented as xi (constant = length of link i)
di = distance Oi-1D positive if oriented as zi-1 (variable if joint i is PRISMATIC)
#i = twist angle between zi-1 and zi around xi (constant)
&i = angle between xi-1 and xi around zi-i (variable if joint i is REVOLUTE)
Robotics 1

13

Denavit-Hartenberg layout made simple


(a popular 3-minute illustration...)

video

https://www.youtube.com/watch?v=rA9tm0gTln8
!

note: the authors of this video use r in place of a, and do not add subscripts!

Robotics 1

14

Ambiguities in defining DH frames


!
!

Robotics 1

frame0: origin and x0 axis are arbitrary


framen: zn axis is not specified (but xn must be
orthogonal to and intersect zn-1)
when zi-1 and zi are parallel: the common
normal is not uniquely defined (Oi can be
chosen arbitrarily along zi)
when zi-1 and zi are incident: the positive
direction of xi can be chosen at will (however,
we often take xi = zi-1 ! zi)
15

Homogeneous transformation

between successive DH frames (from framei-1 to framei)


!

roto-translation around and along zi-1

c&i -s&i 0 0
s&i c&i 0 0
i-1A (q ) =
i
i
0 0 1 0
0 0

0 1

1 0 0 0
0 1 0 0
0 0 1 di =
0 0 0 1

rotational joint ) qi = &i


!

c&i -s&i 0 0
s&i c&i 0 0
0 0 1 di
0 0

0 1

prismatic joint ) qi = di

roto-translation around and along xi


1 0 0 ai
0 c#i -s#i 0
iA =
i
0 s#i c#i 0

Robotics 1

0 0

always a
constant matrix
16

Denavit-Hartenberg matrix
J. Denavit and R.S. Hartenberg, A kinematic notation for lower-pair mechanisms based on matrices,
Trans. ASME J. Applied Mechanics, 23: 215221, 1955

c&i
s&i
i-1A (q ) = i-1A (q ) iA =
i
i
i
i
i
0
0

-c#i s&i
c#i c&i
s#i

s#i s&i
-s#i c&i
c#i

aic&i
ais&i
di

compact notation: c = cos, s = sin


super-compact notation: ci = cos qi, si = sin qi
Robotics 1

17

Direct kinematics of manipulators


yE

RFE

description internal
to the robot using
product 0A1(q1) 1A2(q2)n-1An(qn)
q = (q1,,qn)

slide s
zE

xE

z0

approach a
normal n

RF0

RFB

y0
x0

description external
to the robot using
BTE

BT
E

= BT0 0A1(q1) 1A2(q2) n-1An(qn) nTE

R p
000 1

nsa p
000 1

r = (r1,,rm)

r = fr(q)
Robotics 1

alternative descriptions of the direct kinematics of the robot

18

Example: SCARA robot


video

q3

q1

q2
q4

Sankyo SCARA 8438


Robotics 1

Sankyo SCARA SR 8447


19

Step 1: joint axes

all parallel
(or coincident)

twists # i = 0
or %

J2 elbow

J3 prismatic

J4 revolute

J1 shoulder

Robotics 1

20

Step 2: link axes

the vertical heights


of the link axes
are arbitrary
(for the time being)

a1

Robotics 1

a2

a3 = 0

21

Step 3: frames

z1
x1

axes yi for i > 0


are not shown

z2 = z3
x2

(nor needed; they form


right-handed frames)

x4
z0

x3
z4 = a axis

(approach)

x0
Robotics 1

y0

22

Step 4: DH table of parameters

z1
x1

z2 = z3
x2
x3

x4
z0
x0
Robotics 1

z4
y0

# i$

ai

di

&i

a1

d1

q1

a2

q2

q3

d4

q4

note that:
d1 and d4 could be set = 0
here, it is d4 < 0
23

Step 5: transformation matrices


c&1 - s&1
c&1
0A (q ) = s&1
1 1
0
0
0
0

0 a1c&1
0 a1s&1
1
d1
0
1

c&2 - s&2
c&2
1A (q ) = s&2
2 2
0
0
0
0

0 a2c&2
0 a2s&2
1
0
0
1

q = (q1, q2, q3, q4)


= (&1, &2, d3, &4)
Robotics 1

1
2A (q ) = 0
3 3
0
0
c&4
3A (q ) = s&4
4 4
0
0

0
1
0
0

0
0
1
0

0
0
d3
1

s&4
-c&4
0
0

0
0
-1
0

0
0
d4
1
24

Step 6a: direct kinematics

as homogeneous matrix BTE (products of iAi+1)

c12 -s12
0A (q ,q ,q ) = s12 c12
3 1 2 3
0
0
0
0

0
0
1
0

c4
3A (q ) = s4
4 4
0
0

0
0
-1
0

R(q1,q2,q4)=[ n s a ]

s4
-c4
0
0

c124 s124 0
0
BT = 0A (q ,q ,q ,q ) = s124 -c124
E
4 1 2 3 4
0
0
-1
(BT0 = 4TE = I)
0
0
0
Robotics 1

a1c1+ a2c12
a1s1+ a2s12
d1+q3
1
0
0
d4
1

p = p(q1,q2,q3)

a1c1+ a2c12
a1s1+ a2s12
d1+q3+d4
1
25

Step 6b: direct kinematics


as task vector r IRm

c124 s124 0
0
0A (q ,q ,q ,q ) = s124 -c124
4 1 2 3 4
0
0
-1
0
0
0
extract #z
from
R(q1,q2,q4)

Robotics 1

a1c1+ a2c12
a1s1+ a2s12
d1+q3+d4
1

px
a1c1+ a2c12
r = py = fr(q) = a1s1+ a2s12 IR4
pz
d1+q3+d4
#z
q1+q2+q4

take
p(q1,q2,q3)
as such

26

Stanford manipulator
!

6-dof: 2R-1P-3R (spherical wrist)


!
!

shoulder offset
one possible DH assignment of
frames is shown
determine the associated
!
!

write a program for computing


the direct kinematics
!
!

Robotics 1

DH parameters table
homogeneous transformation
matrices
direct kinematics

numerically (Matlab)
symbolically (Mathematica,
Maple, Symbolic Manipulation
Toolbox of Matlab, )
27

DH table for Stanford manipulator


!

6-dof: 2R-1P-3R (spherical wrist)

ai

di

-!/2

d1>0

q1=0

!/2

d2>0

q2=0

q3>0

-!/2

-!/2

q4=0

!/2

q5=-!/2

d6>0

q6=0

joint variables are in red,


with their current value in the shown configuration
Robotics 1

28

KUKA KR5 Sixx R650


!

6R (offsets at shoulder and elbow, spherical wrist)

top view
!

determine
!
!
!

Robotics 1

side view (from an observer in V)

frames and table of DH parameters


homogeneous transformation matrices
direct kinematics

available at
DIAG Robotics Lab
29

KUKA LWR 4+
!

7R (no offsets, spherical shoulder and spherical wrist)


available at
DIAG Robotics Lab
!

determine
!

!
!

front
Robotics 1

front view

frames and table


of DH parameters
homogeneous
transformation
matrices
direct kinematics
d1 and d7 can be
set = 0 or not
(as needed)

back
side view

30

Appendix:
Modified DH convention
!

a modified version used in J. Craigs book Introduction to Robotics, 1986


!
!
!
!
!

has zi axis on joint i


ai & #i = distance & twist angle from zi to zi+1, measured along & about xi
di & &i = distance & angle from xi-1 to xi, measured along & about zi
source of much confusion... if you are not aware of it (or dont mention it!)
convenient with link flexibility: a rigid frame at the base, another at the tip...
y2

classical % c# i
i"1

' s# i
Ai = '
0
'
& 0

"c$ i s# i
c$ i c# i
s$ i
0

s$ i s# i
"s$ i c# i
c$ i
0

ai c# i (
ai s# i *
di *
*
1 )

planar 2R
example

y1

y0

x2

x1

x0

modified %

i"1

Aimod

c# i
'c$ i"1 s# i
='
s$ s#
' i"1 i
& 0

Robotics 1

"s# i
c$ i"1 c# i
s$ i"1 c# i
0

0
ai"1 (
"s$ i"1 "di s$ i"1 *
c$ i"1
di c$ i"1 *
*
0
1 )

y1

y0

y2

x2

x1
x0

modified DH tends to place frames


at the base of each link

31

You might also like