Professional Documents
Culture Documents
Feedback Example:
The Inverted Pendulum
26-1
Signals and Systems
26-2
toward each other, meeting at the origin and then traveling along the imagi-
nary axis. The presence of a pole in the right half-plane indicates that even
with a very small input (such as a small displacement of the rod), the rod angle
will increase exponentially. With the poles on the imaginary axis, any dis-
placement will result in an oscillatory behavior. Consequently, the system re-
mains unstable for all values of the feedback gain.
A second type of feedback to consider corresponds to choosing the cart
acceleration proportional to the derivative of the angular displacement. This
choice is motivated by the possibility that perhaps the pendulum can be stabi-
lized by accelerating the cart faster if the angular displacement is increasing
faster. Examination of the root locus with derivative feedback demonstrates
again that for the feedback gain less than zero the system becomes increas-
ingly unstable while with the feedback gain greater than zero the system be-
comes more stable but is never completely stabilized. Finally, we consider a
combination of proportional plus derivative feedback, and in this case by ap-
propriate choice of the two gain factors the system can be stabilized.
The demonstration accompanying this lecture, besides being entertain-
ing, is hopefully very instructive. In addition to showing that the system can in
fact be stabilized by appropriate choice of feedback, we are able to show that
feedback is able to compensate for external disturbances and to changes in
the system dynamics.
Suggested Reading
Section 11.3, Root-Locus Analysis of Linear Feedback Systems, pages 700-
704
Feedback Example: The Inverted Pendulum
26-3
DEMONSTRATION
26.1
A simple inverted
pendulum.
'NI
N
x(t) TRANSPARENCY
L 26.1
The inverted
pendulum and a
representation of the
open-loop system.
External
disturbances
x(t) 1 System
a(t) p dynamics
Signals and Systems
26-4
TRANSPARENCY
26.2
The inverted
pendulum with a
feedback loop from
the pendulum angle to
the cart acceleration.
External
disturbances
L I
TRANSPARENCY
26.3
Equations describing
the dynamics of the
inverted pendulum.
a(t)
s(t)
d 2 6(t)
L d2 =g sin [6(t)] + Lx(t) - a(t) cos 0(t)
cos [O(t)] 1
Feedback Example: The Inverted Pendulum
26-5
2
L d 6(t
t - g 0(t) = Lx(t) - a(t)
dt 2
X X Re
DEMONSTRATION
26.2
The cart and inverted
pendulum represented
by Transparency 26.2.
Signals and Systems
26-6
L +
x(t) 0(t)
a (t)
H(s) =
TRANSPARENCY Ls2 _ g
26.6
System function and
block diagram with
feedback.
L +
x(t) 0(t)
a t)
LH(s)
O(s) = + (s) X (s)
1 + G (s) H(s)
Feedback Example: The Inverted Pendulum
26-7
TRANSPARENCY
Ls 2 g 26.7
System function with
proportional feedback.
Proportional Feedback:
a(t) = K 1 0(t)
G(s) = K3
=) 1
((s) X(s)
2 _ (g - K,3
g -K31
E-K
Poles at s = +
1L:7
Proportional Feedback:
O(s) = 1 X(S)
S2 _ TRANSPARENCY
LK1 26.8
Locus of system
function poles for
Im proportional feedback
with ki < 0.
s-plane
K1 <0
4 XRe
-vg /L + \/gji -
Signals and Systems
26-8
Proportional Feedback:
(s)= 1 X(s)
TRANSPARENCY s2 _
26.9
Locus of system
function poles for
proportional feedback Im
with ki > 0.
s-plane
K1> 0
A 0 A 4( XRe
-v I +v'97i
L H(s)
0(s) = H(s) X(s)
1 + G(s) H(s)
H (s) = s1_
TRANSPARENCY Ls2 _ g
26.10
Overall system
function with
derivative feedback. Derivative feedback:
a(t) = K2 dO(t)
dt
G(s) = K2 s
8()= 1 X(s)
s2 + s(K 2 /L) - g/L
Poles at
K2
(K)2(
Feedback Example: The Inverted Pendulum
26-9
Derivative feedback:
0(s) = X(s)
S2 + s(K 2 /L) - g/L
22 TRANSPARENCY
26.11
K2 + Locus of system
2 (.\
2L - \ 2Li \L/ function poles for
derivative feedback
with k2 < 0-
s-plane
K2 <0
X---O-- Re
- 1/9-/L
Derivative feedback:
0(s) =
S2 + s(K 1X(s)
2 /L) - g/L
TRANSPARENCY
26.12
29 K2 ( \ Locus of system
s -- +(L + \/ function poles for
proportional feedback
with k2 > 0.
K2>0 s-plane
Re
-/vL+ /g-/
Signals and Systems
26-10
H(s) =
s2 + s (K2 /L) - g/L + K1 /L
Poles at
-K 2 (K2 2 Kj -9
s 2L V 2L L
H(s) =
TRANSPARENCY
s2 + s (K2 /L) - g/L + K, /L
26.14
Locus of system -K2 K2 2 K -g
function poles for = 2L V(\ )/~
proportional plus
derivative feedback
with k2 > 0 and ki > 0.
s-plane
K2 >0
K1 >0
Re
Feedback Example: The Inverted Pendulum
26-11
MIT OpenCourseWare
http://ocw.mit.edu
The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.