Professional Documents
Culture Documents
Reading assignment: Ch 7, sections 7.1, 7.5, and 7.7.1 (we previously covered concepts in sections 7.2, 7.3, 7.4, and
7.6)
Laboratory sections meet during the week of December 10.
Problem 40: (25 points)
Consider the closed-loop system in Figure 1 where
1
Gp (s) = .
s(10s + 1)
1. An analysis showing how the controller parameters are chosen, including appropriate Bode plots.
2. A Bode magnitude and phase plot of the compensated open-loop transfer function Gc(s)Gp (s). Indicate the
gain-crossover frequency and phase margin on the Bode plot.
3. A plot of the unit-step response of the close-loop system. Use MATLAB to determine the peak overshoot.
4. If the peak overshoot is not within specification, a brief description of what parameter(s) of the design you
would change in the next iteration of the control design and the reason for doing so.
If all the state variables are accessible, then it is possible to use full-state feedback
where K is a 1 by n row vector of constant gains, g is a scalar multiplier, and r(t) is a scalar input to the closed-loop
system. The controllability matrix P is defined as
P = G, F G, F 2G, . . ., F n−1G .
If the matrix P is nonsingular, then the constant gain vector K can be chosen to place the poles of the closed-loop
transfer function Y (s)/R(s) at any desired location. The proof of this result is beyond the scope of EE 428, but it
is rigorously proven in the graduate controls course EE 527.
1. (4 points) Suppose that the control law
u(t) = Kx(t) + gr(t)
is applied to the plant
Now consider the open-loop system described by the state space model
0 1 0
F = ,G = ,H = 1 1 .
−8 6 3
9. (2 points) For the parameters chosen in part 8, specify the zero location of the closed-loop system. Does
full-state feedback affect the location of the open-loop zero?
Suppose that the state vector x(t) is not available for measurement, and so we can not directly implement the control
law in part 6.
10. (3 points) Determine the observability matrix Q and its rank. Is the open-loop system observable?
11. (5 points) Estimate the state-vector x(t) as x̂(t) using the full-order observer
˙
x̂(t) = F x̂(t) + Gu(t) + L(ŷ(t) − y(t))
ŷ(t) = H̄ x̂(t),
B(s) 3s + 3
Gp (s) = = 2 .
A(s) s − 6s + 8
1. (20 points) Suppose that the desired response of the closed-loop system is specified by the transfer function
Ym (s) Bm (s) 2
= = 2 .
Uc (s) Am (s) s + 2s + 2