Professional Documents
Culture Documents
1
Jizhong Xiao
Department of Electrical Engineering
City College of New York
jxiao@ccny.cuny.edu
Manipulator Control
Introduction to ROBOTICS
The City College of New York
2
Outline
Homework Highlights
Robot Manipulator Control
Control Theory Review
Joint-level PD Control
Computed Torque Method
Non-linear Feedback Control
Midterm Exam Scope
The City College of New York
3
Homework 2
+ s s
s s
+ s s
+ + s s
= =
y and x f or
y and x f or
y and x f or
y and x f or
x y a
0 90
90 180
180 90
90 0
) , ( 2 tan
u
u
u
u
u
2
1
sin = x
Joint variables ?
Find the forward kinematics, Roll-Pitch-Yaw
representation of orientation
Why use atan2 function?
Inverse trigonometric functions have multiple solutions:
? = x
2
3
cos = x
? = x
150 , 30 ) 2 / 1 ( sin
1
=
30 , 30 ) 2 / 3 ( cos
1
=
30 ) 2 / 3 , 2 / 1 tan( = a
) tan( ) tan( t k x x + = Limit x to [-180, 180] degree
The City College of New York
4
Homework 3
Z1
x1
y1
x2
y2
Z2
m2
m1
1
u
2
u
L
L
(
(
(
(
+
+
= =
1 0 0 0
0 1 0 0
12 1 0 12 12
12 1 0 12 12
2
1
1
0
2
0
S S C S
C C S C
T T T
(
(
(
(
=
1 0 0 0
2
0
z z z z
y y y y
x x x x
p a s n
p a s n
p a s n
T
Find kinematics model of 2-link robot, Find the inverse kinematics solution
)
`
= +
= +
y
x
n
n
) sin(
) cos(
2 1
2 1
u u
u u
)
`
= + +
= + +
y
x
p
p
1 2 1
1 2 1
sin ) sin(
cos ) cos(
u u u
u u u
Inverse: know position (Px,Py,Pz) and
orientation (n, s, a), solve joint variables.
) , ( 2 tan
2 1 x y
n n a = +u u
) , ( 2 tan
1 x x y y
n p n p a = u
The City College of New York
5
Homework 4
Z1
x1
y1
x2
y2
Z2
m2
m1
1
u
2
u
L
L
Find the dynamic model of 2-link
robot with mass equally distributed
) ( ) , ( ) ( q C q q H q q D + + = t
=
=
n
k i j
T
ji j jk ik
U J U Tr D
) , max(
) (
>
s
=
c
c
i j for
i j for T Q T
q
T
U
i
j j
j
j
i
ij
0
1
1
0 0
(
+
(
+
(
=
(
) (
) (
) , (
) , (
2
1
2
1
2
1
22 21
12 11
2
1
u
u
u u
u u
u
u
t
t
c
c
h
h
D D
D D
Calculate D, H, C terms directly
Physical meaning?
< <
> >
> >
=
c
c
k i or j i
k j i T Q T Q T
j k i T Q T Q T
U
q
U
i
j j
j
k k
k
i
k k
k
j j
j
ijk
k
ij
0
1
1
1
1
0
1
1
1
1
0
j
j ji
n
i j
j i
r U g m C
=
=
=
=
n
m k i j
T
ji j jkm ikm
U J U Tr h
) , , max(
) (
Interaction effects of motion of joints j & k on link i
The City College of New York
6
Homework 4
Z1
x1
y1
x2
y2
Z2
m2
m1
1
u
2
u
L
L
Find the dynamic model of 2-link
robot with mass equally distributed
i
i i
q
L
q
L
dt
d
t =
c
c
c
c
) (
P K L =
) ( ) , ( ) ( q C q q H q q D + + = t
Derivation of L-E Formula
(
(
(
(
=
1
i
i
i
i
i
z
y
x
r
(
= =
}
}
= =
i
p
r p
T
ir
T
i
i
i
i ip
i
r
i i
q q U dm r r U Tr dK K
1 1
) (
2
1
= =
=
c
c
= =
i
j
i
i j ij
i
i
i
j
j
j
i
i i
i
r q U r q
q
T
r
dt
d
V V
1 1
0
0 0
) ( ) (
point at link i
Velocity of point
Kinetic energy of link i
i
J
Erroneous answer
(
(
(
(
=
1
0
0
L
r
i
i
The City College of New York
7
Homework 4
X0
Y0
X1
Y1
1
u
L
m
(
(
(
(
=
1
0
0
1
1
l
r
Example: 1-link robot with point mass (m)
concentrated at the end of the arm.
Set up coordinate frame as in the figure
2
1
2
2
1
u
m l K =
1
8 . 9 u S l m P =
P K L =
1 1
2
1 1
8 . 9 ) ( u u
u u
t C l m m l
L L
dt
d
=
c
c
c
c
=
=
n
t
t
t
1
Driving torque
applied on each link
Non-linear, highly coupled , second order differential equation
Joint torque Robot motion
The City College of New York
10
Jacobian Matrix Revisit
4 4
6
0
1 0 0 0
=
p a s n
T
(
(
(
=
(
(
(
=
) (
) (
) (
3
2
1
q h
q h
q h
z
y
x
p
(
(
(
=
(
(
(
) (
) (
) (
) (
) (
) (
} , , {
6
5
4
q h
q h
q h
q
q
q
a s n
u
|
Forward Kinematics
(
(
(
(
= =
) (
) (
) (
) (
6
2
1
1 6
q h
q h
q h
q h Y
1 6 1 6
=
n n
q J Y
(
(
(
(
(
(
(
(
u
|
z
y
x
Y
1
2
1
6
) (
(
(
(
(
=
n
n
n
q
q
q
dq
q dh
n
n
n
n
n
q
h
q
h
q
h
q
h
q
h
q
h
q
h
q
h
q
h
dq
q dh
J
(
(
(
(
(
(
(
(
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
=
|
|
.
|
\
|
=
6
6
2
6
1
6
2
2
2
1
2
1
2
1
1
1
6
) (
=
(
+ +
+ +
=
(
) , (
) , (
) sin( sin
) cos( cos
2 1 2
2 1 1
2 1 2 1 1
2 1 2 1 1
u u
u u
u u u
u u u
h
h
l l
l l
y
x
(
+ + +
+ +
=
(
(
(
(
c
c
c
c
c
c
c
c
=
) cos( ) cos( cos
) sin( ) sin( sin
2 1 2 2 1 2 1 1
2 1 2 2 1 2 1 1
2
2
1
2
2
1
1
1
u u u u u
u u u u u
u u
u u
l l l
l l l
h h
h h
J
(
=
(
=
2
1
u
u
J
y
x
Y
The City College of New York
12
Robot Manipulator Control
Joint Level Controller
=
= + +
) (
) ( ) , ( ) (
q h Y
q C q q H q q D t
Task Level Controller
Find a control input (tor),
t as q q
d
t as Y Y
d
0 = Y Y e
d
Find a control input (tor),
Robot Controller
_
tor
q q
d
q
d
q
d
q
Trajectory
Planner
q q e
d
=
e e
Task level
Planner
Robot
Dynamics
Controller
_
tor q q
d
Y
d
Y
Forward
Kinematics
Y Y
Y Y e
d
=
e e
Robot System:
The City College of New York
13
Robot Manipulator Control
Control Methods
Conventional Joint PID Control
Widely used in industry
Advanced Control Approaches
Computed torque approach
Nonlinear feedback
Adaptive control
Variable structure control
.
The City College of New York
14
Control Theory Review (I)
actual
a
desired
d
V
Motor
actual
a
-
compute V using
PID feedback
d
a
Error signal e
PID controller: Proportional / Integral / Derivative control
e=
d
a
V = K
p
e + K
i
e dt + K
d
)
d e
dt
Closed Loop Feedback Control
Reference book: Modern Control Engineering, Katsuhiko Ogata, ISBN0-13-060907-2
The City College of New York
15
Evaluating the response
How can we eliminate
the steady-state error?
steady-state error
settling time
rise time
overshoot
overshoot -- % of final value
exceeded at first oscillation
rise time -- time to span from
10% to 90% of the final value
settling time -- time to reach
within 2% of the final value
ss error -- difference from the
systems desired value
The City College of New York
16
Control Performance, P-type
K
p
= 20
K
p
= 200
K
p
= 50
K
p
= 500
The City College of New York
17
Control Performance, PI - type
K
p
= 100
K
i
= 50 K
i
= 200
The City College of New York
18
Youve been integrated...
K
p
= 100
unstable &
oscillation
The City College of New York
19
Control Performance, PID-type
K
p
= 100
K
i
= 200 K
d
= 2
K
d
= 10 K
d
= 20
K
d
= 5
The City College of New York
20
PID final control
The City College of New York
21
Control Theory Review (II)
Linear Control System
State space equation of a system
Example: a system:
Eigenvalue of A are the root of characteristic equation
Asymptotically stable all eigenvalues of A have negative
real part
Bu Ax x + =
=
=
u x
x x
2
2 1
u
x
x
x
x
(
+
(
=
(
1
0
0 0
1 0
2
1
2
1
0 = A I
0
0
1
2
= =
A I
(Equ. 1)
The City College of New York
22
Control Theory Review (II)
Find a state feedback control such that the
closed loop system is asymptotically stable
Closed loop system becomes
Chose K, such that all eigenvalues of A=(A-BK) have
negative real parts
}
A
B
-K
x
x
u
x K u =
| |
(
=
2
1
2 1
x
x
k k u
x BK A x ) ( =
(Equ. 2)
0
1
'
1 2
2
2 1
= + + =
+
= k k
k k
A I
] ) ( ) ( ) ( [
1 1
V x G x f x G U
+ =
V X =
Dynamic
System
U
x
Nonlinear
Feedback
Linear System
V
U x x = +cos
V x U + = cos V x =
Original system:
Nonlinear feedback:
Linear system:
The City College of New York
24
Robot Motion Control (I)
Joint level PID control
each joint is a servo-mechanism
adopted widely in industrial robot
neglect dynamic behavior of whole arm
degraded control performance especially in
high speed
performance depends on configuration
Robot Controller
_
tor
q q
d
q
d
q
d
q
Trajectory
Planner
q q e
d
=
e e
The City College of New York
25
Robot Motion Control (II)
Computed torque method
Robot system:
Controller:
=
= + +
) (
) ( ) , ( ) (
q h Y
q C q q H q q D t
) ( ) , ( )] ( ) ( )[ ( q C q q H q q k q q k q q D tor
d
p
d
v
d
+ + + + =
0 ) ( ) ( ) ( = + + q q k q q k q q
d
p
d
v
d
0 = + + e k e k e
p v
Error dynamics
How to chose
Kp, Kv ?
Advantage: compensated for the dynamic effects
Condition: robot dynamic model is known
The City College of New York
26
Robot Motion Control (II)
0 = + + e k e k e
p v
e x
e x
=
=
2
1
1 2 2
2 1
x k x k x
x x
p v
=
=
=
(
2
1
2
1
1 0
0
1
2
= + + =
+
=
p v
v p
k k
k k
A I
2
4
2
2 , 1
p v v
k k k
=
In matrix form:
Characteristic equation:
The eigenvalue of A matrix is:
Condition: have negative real part
0 >
v
k
0 >
p
k
One of a
selections:
The City College of New York
27
Robot Motion Control (III)
=
= + +
) (
) ( ) , ( ) (
q h Y
q C q q H q q D t
q J q q h
dq
d
Y
= = )] ( [
Non-linear Feedback Control
q J q J Y
+ = ) (
1
q J Y J q
=
t = + +
) ( ) , ( ) ( ) (
1
q C q q H q J Y J q D
Task level
Planner
Robot
Dynamics
Linear
Controller
_
U
q
q
d
Y
d
Y
Forward
Kinematics
Y
Y
Y Y e
d
=
e
e Nonlinear
Feedback
tor
Linear System
Jocobian:
Robot System:
The City College of New York
28
Robot Motion Control (III)
Task level
Planner
Robot
Dynamics
Linear
Controller
_
U
q
q
d
Y
d
Y
Forward
Kinematics
Y
Y
Y Y e
d
=
e
e Nonlinear
Feedback
tor
Linear System
) ( ) , ( ) ( ) (
1
q C q q H q J U J q D tor + + =
U J q D Y J q D
1 1
) ( ) (
=
Design the nonlinear feedback controller as:
Non-linear Feedback Control
Then the linearized dynamic model:
U Y =
Design the linear controller:
) ( ) ( Y Y k Y Y k Y U
d p d v d
+ + =
Error dynamic equation: 0 = + + e k e k e
p v
The City College of New York
29
Project
Task level
Planner
Robot
Dynamics
Linear
Controller
_
U
q
q
d
Y
d
Y
Forward
Kinematics
Y
Y
Y Y e
d
=
e
e Nonlinear
Feedback
tor
Linear System
Simulation study of Non-linear Feedback Control
The City College of New York
30
Thank you!
x
y
z
x
y
z
x
y
z
x
z
y
HWK 5 posted on the web, Due: Oct. 23 before class
Next Class (Oct. 23): Midterm Exam
Time: 6:30-9:00, Please on time!