This action might not be possible to undo. Are you sure you want to continue?

Welcome to Scribd! Start your free trial and access books, documents and more.

Find out moreThird Edition

Solutions Manual

Karl J. Åström Björn Wittenmark

Department of Automatic Control Lund Institute of Technology October 1997

Preface

This Solutions Manual contains solutions to most of the problems in the fourth edition of str m, K. J. and B. Wittenmark 1997: Computer controlled Systems Theory and Applications, Prentice Hall Inc., Englewood Cli s, N. J. Many of the problems are intentionally made such that the students have to use a simulation program to verify the analytical solutions. This is important since it gives a feeling for the relation between the pulse transfer function and the time domain. In the book and the solutions we have used Matlab Simulink. Information about macros used to generate the illustrations can be obtained by writing to us. It is also important that a course in digital control includes laboratory exercises. The contents in the laboratory experiments are of course dependent on the available equipment. Examples of experiments are ? Illustration of aliasing ? Comparison between continuous time and discrete time controllers ? State feedback control. Redesign of continuous time controllers as well as controllers based on discrete time synthesis ? Controllers based on input-output design ? Control of systems subject to stochastic disturbances. Finally we would like to thank collegues and students who have helped us to test the book and the solutions. Karl J. str m Bj rn Wittenmark

Department of Automatic Control Lund Institute of Technology Box 118 S-220 00 Lund, Sweden

i

Solutions to Chapter 2

Problem 2.1

The system is described by

x = ,ax + bu _ y = cx Sampling the system using 2.4 and 2.5 gives

,

b xkh + h = e,ah xkh + 1 , e,ah a ukh y kh = cxkh The pole of the sampled system is exp,ah. The pole is real. For small values of h the pole is close to 1. If a 0 then the pole moves towards the origin when h increases. If a 0 then the pole moves along the positive real axis.

Problem 2.2

a. Using the Laplace transform method we nd that

**,,sI , A,1 = L,1 1 8 s = eAh = L,1 s2 + 1 : ,1 8 cos h sin h 9 = : , sin h cos h ; 8 1 , cos h 9 Zh Zh 8 sin s 9 As B ds = : ,= e cos s ; ds = : sin h ;
**

0 0

1 = s;

9!

**b. The system has the transfer function + 2 1 s Gs = s2 + +s3+ 2 = s +s1s3+ 2 = s + 1 , s + 2 3 Using the Table 2.1 gives
**

,2h e,h H q = 2 1 , e,h , ,1 , e ,2h q, 2 q,e

c. One state space realization of the system is 80 0 09 819 x = :1 0 0;x + :0;u _ 0 80 1 09 y = :0 0 1;x 1

0:5y = 6u xkh + h = 0:5xkh + 6ukh xt = axt + but _ ykh = xkh y t = xt discrete-time system continuous time system ah = ln 0:5 a = . h =6 3 q2 q = h 6q+ 41+ 1 . 12 . = q .3 = eAh a. .1. . 0:5q .= and Zh 0 8 1 0 09 eAh = I + Ah + A2 h2 =2 + : : : = : h 1 0 . h2=2 h 1 eAs B ds = 0 Zh 8 9 8 9 : 1 s s2=2 . 2 . 1 = 6uk . 1 y . 1 = 6 0 ln b = eah6a 1 = 12h 2 . 0:5yk . 13 8 x x x 98 h 9 2 x x : h2q + 1=2 hq x 1 q .T 8 9 :0 0 1. =0 easb ds = 6 Zh 0 b h b eas b ds = a eas = a eah . : h3=2 .1 u qy .T ds = : h h2=2 h3=6 .1 y = 6q . A3 = 0 0 . ln 2 h 8 =eah = 0:5 Rh : . 1 A = h ln yk . H q =C qI .Now then 80 0 A2 = : 0 0 1 0 09 0 . 3 Problem 2.

8 . 1. a.:1 q . xkh + : 0:7 . 1 + + k . 8 s + 0:5 . A.1 9.b. = : 1 9 8 q . No equivalent continuous system exists.h = : sin h .1 1 .1 9 detsI .x Pulse transfer operator 8 9 H q = C qI . yk + 0:5y k .3 and 3.0:3 c. Problem 2. yk + 1 = .0:5 2 = . h sin xk + 1 = xk + .1 . cos h cos h . 8 9819 q + 1 1 81 09 q 1 . : 1 .0:5 1 9 8 0:5 9 xkh + h = : 0 . .1 q . 2 0 1 8 9 8 9 8 9 y = :1 0. = q 2 + 1 8 z 1 Y z = H zU z = zz2 + 11 z . :1.0:3 . = : 0 s + 0:3 .2a. sin y k = Cxk 8 1 . 1 = 6uk . 8 Eigenvalue to : 1 = . = 0 .1 8 1 9 0.4 Harmonic oscillator cf. s + 0:5s + 0:3 = 0 Both eigenvalues of on the negative real axis No corresponding continuous system exists. = : 1 . ukh : ykh = 8 1 1 9 xkh : . 1 = z2 1 1 + z .0:5y k + 6ut H q = q +60:5 one pole on the negative real axis. . cos h 9 h .: = q2 + 1 : .uk h = : . 1 . 0 h = = : .

k . 1 + .

k . cos k . 1 = 1 .

1 yk = sin 2 2 3 .k .

where .

36 q . cos t. . . 1 + 1 1 . khukhds Problem 2.1 8 0 9 1 Gs = see Probl. 2.2 = : 1 0 . 36 1 + 1 s + 2 .2 1 .3 8q e q e Problem 2.7 The representation 2. s2 s 1 s + 8 9 8 s .1: 1 5 1 1 Gs = s2 s + 2s + 3 = 1 s12 . = s2 + 1 b. 27 1 .2 . 1 s + 3 6 s 4 9 5 1 . y t = 1 . 1 is a step at k = 1.ukh ykh = Cxkh and the controllable form realization is ~ ~ z kh + h = z kh + .3 1 +1 H q = 12 qq.k .6 Integrating the system equation x = Ax + Bu _ gives xkh + h = eAh xkh + kh+h Z kh Ah xkh + Bukh =e eAs B s . e. cos k 2 Y s = ss21+ 1 = 1 . and y kh = 1 . : 1 s .7 is xkh + h = xkh + . e. cos t y kh = 1 . cos k Notice 4 that the step responses of the continuous time and the zero-order hold sampled systems are the same in the sampling points. 12 . The same way as a. y t = 1 .ukh ~ ykh = Cz kh 4 . Problem 2.1 9. : 1 .5 Do a partial fraction decomposition and sample each term using Table 2.

1 9 0. = .t22 = 1=2h or 1 h T = 2h2 : 2 . 0::5 z :5 Problem 2.8 The pulse transfer function is given by 8 98 : 1 0 . 9 8 h2=2 9 8 9 ~ = : 2 .=: h . ~ = T T . 0 1 = z2z . . u = Ax + Bu _ 5 .where 81 = :0 8 C = :1 h.9 The system is described by 8 . 1 0 8 2 2 9 ~ C = : h =2 h =2 . ~ C = CT .1 .1 or or ~ T = T ~ CT = C ~ . t21 t21 = t11 ht11 + t12 = 2t12 . zz .a b 9 8 f 9 x = : c . z .1 ~ . : 1 . t22 ht21 + t22 = t12 h2 t + ht = 1 12 2 11 h2 t + ht = 0 22 2 21 h2 t + h2 t = 1 2 11 2 21 h2 t + h2 t = 0 2 12 2 22 Equations 1 4 now give 1 2 3 4 t11 = t21 = 1=h2 t12 = . = T. 0:2 9 8 2 9 .0 2z . = :1.1. 0:5 .h . 0:2 = z z .d . = H z = C zI . x + : g . 0 8 9 From Section 2. 2 8 9 Problem 2.5 we get This gives the following relations t11 = 2t11 . 0:5 : 0 z .

d2 + 4bc =. 2e 0 1 2 0 1 1 bh =: . There are multiple eigenvalues if both a = d and bc = 0. b. we notice that 0= 1= 1 2 Zh 0 0 and 0 sds = . 1 dh e1h . Using the result in Appendix B we nd that eAh = 0I + 1Ah and e1h = 0 + 11h e2h = 0 + 12h This gives 2 h . 2 2 1 1 1 . 1 and 2. e2h . e2h 1 = . 1 1 .The eigenvalues of A is obtained from + a + d . 1 ch 0 . 1 h 8 . h To compute . bc = 0 which gives r a + d a . c. 2 e1h . ah 9 = 1e . and d are nonnegative implies that the eigenvalues. bc = 2 + a + d + ad . 2 4 The condition that a. 1 . are real.

1 depend on h. 1 2 1 . 1 . Introduce Zh 0 s 1 sds = .

= 0 B + 1AB Problem 2. e1 h . 1 . Using the result from Problem 2. 1 2 then .10 a.0:0197 e1 h = 0:7895 2 = .9 gives 1 = . 6 8 9 . = 0 I + 1A B = : 00::0296 .0:0129 e2 h = 0:8566 Further 0 = 0:9839 1 = 0:8223 and 8 0:790 0 9 = 0I + 12 1 A = : 0:176 0:857 . 0 = 11:9412 1 = 63:3824 281 . 1 .1.e2 h .

11 The motor has the transfer function Gs = 1=ss + 1.h . 1 .1 8 9 C = :0 1. 1 .h z . 1 1 . The pulse transfer operator is now given by H q= C qI .0:176 q .h he +e z+ 8 9 8 z . he .1 + e.1 s 0 1 ss + 1 : 1 s + 1 . 0:790 0 . e. 1 0 = expAh = L.2.h =: 1 . .: 1 . which agrees with the pulse transfer operator given in the problem formulation. The pulse transfer function is H z = C zI .s . A state space representation is given in Example A.s 9 : 1 . Problem 2. .1 0 . .h 0 9. = : 0 0 030 = q 2 :.h 9 :0 1. 1 . e.h . e.1 .h = h + e z 2 . = 8 9! eAs Bds = 8 1 . e1 .h z .1 8 1 .2.h 9 = :0 1. 1 . e.b. A.h + = h + e z 1ze. 1 0 9 = L. where A = : .h . 1:q + 0:026 65q + 0:68 8 9. : h + e. . 8 9 8 9 8 1 . k = 1 C k Now k = eAh k = eAkh 7 . 1 z + . . This gives the sampled representation given in Example A.h .18 0:281 9 98 q .= 8 e. ds = : h + e. e.h 9 Zh 8 e. 0:857 . e. .h z . = .1 . : h + e. . a.: .1 8 9 B = :1.h .h1z. : 0:0297 .1.h Zh 0 0. e. 8 9 . z. The pulse response is 0 hk = 0 k.1 0 = z .h .h . 1 : 1 . e. b. e.h .h . 0 sI . e e.

_ 8 . = 1 .h ukh d.k.1h 1 . 1ukh + h + 1 .s s h = 1.1 The pole solid and zero dashed in Problem 2. 1 = 8 = h .h . e. Gs = 1 e.h z = .kh An alternative way to nd hk is to take the inverse z -transform of H z .11. = : 0 9 8 e.h 9 1.h. e.12 Consider the following transfer operators of in nity order. = 0 x + 1 ut . e.1 to 0 when h goes from zero to in nity.: .1h + e. : h + e. = 0:5 h in nite order x = ut . he. 1 . 2. . = 1 .1 . The poles are in z = 1 and z = exp.k. A di erence equation is obtained from H q ykh + 2h . e. e.1h + h + e.1 Pole and zero 0 −1 0 Figure 2. 1 = f h h h The function f h goes from . Problem 2. 10 Sampling period h 20 This gives C k. c.h .h he. 1 + e.h ykh + h + e.h y kh = = h + e.k. The zero is in e.h . e. + e..1.1h 0 9 8 1 .h 1 . See Fig.d as function of the sampling period.k. . 1 .h . The second pole will move from 1 to the origin when h goes from zero to in nity.

Z 0 eAs ds B = Z0:5 0 ds = 0:5 xk + 1 = xk + 0:5uk + 0:5uk .0uk + .1 Problem 2.uk 9 k 0 . The pulse-transfer function is 8 9 . 2 9 .8 uk .1 z .= zz .1 .1 : H z =C zI . 1 z : + 1 z z H z = 0z5z.13 a. 1 Notice that = eAh = e0 = 1 .1 h.6 but with = 1:5.a. . 1 + . 1 z | z | z k=0 hkh = 0:5 k = 1 1 k 1 8 9 8 9 Invers transform of H z using Table 2. : z 0:5 9 0:5 . 1 . . This is the same system as in Example 2. 1 = 0:5 z . Discrete time system is given by cf. 1 . 1 . : z . k1 =1. k 98 9 8 9 The system is of second order. Z 0 h eAs ds B = Z0:5 0 1 ds 1 = 0:5 = eAh. b. : 015 . = : 1 0 . 0 z 8 9 8 8 9 1 8 z 0:5 9 0:5 9 1 8 :1 0.3 gives ! 0 =1. k2 c. 19 + : 1 . uk= xk+ . 1 + z . CCS p.1 uk . 38 40. = : :1 z z . 1 : + 1 = 0z5z.0 = . z = 1 and the zeros: z = .0uk . : ux. In this case we have d = 2 and 0 = 0:5 and we get xk + 1 = xk + .1 uk .0:5 . 1 State space model: 8 xk + 1 9 8 1 : uk .1. 1: 0 z . The poles are z = 0. xk + 1 = xk + . = : 0 8 y k = : 1 0:5 : xk . 0:5 0 .2 z .

e.0:5 .0 .5 + 0:02z.1 = e. c.1 0 z .3 k 3 .14 Sampling the given di erential equation using the same procedure as in Problem 2. 1 1 8 xk 9 8 9 yk = : 1 0 0 . : 0 . uk .4 + 0:05z.3 + 0:14z. e. 2 .2 + .0 = 1 . where Problem 2.6 + : : : The pulse response is now given by the coe cients in the series expansion.k. sb ds = b 1 . = e.1 . h Using the de nition of 0 and with h = 1 it follows that d = 4 and = 3+ 0 Further h. h. e.k.0:5 = 0:39 .1 .1 .4 and the de nition of the z -transform 2.0 9 8 xk 9 8 0 9 k : uu. The pulse transfer function is 9 8 9 8 8 9 z . + : 0 . h .0:62. s b ds b .13 gives a = e. = 10 0 Z0 0 e. 0037 z .1 = 0:24 A state representation is obtained from 2. : uk .0 e Using Theorem 2. : 0 1 0 0 z + :24 .27 we can also get the pulse response by expanding H z in z .1 . z z : Some calculations give 80 k1 hk = : . .1 e. 2 . h.0 k=2 .. i. : uk . 0 Z b = e. e. H z = 0:39z. 1 b. h. = : 0 0 1 . uk k 0 0 0 uk . There are three poles p1 = p2 = 0 and p3 = 0:37 and one zero z1 = .. = z 20z + .1 = 0:239z. 0 e .1 = 0:37 . 0 3 0 b4 = e. e.12 8 xk + 1 9 8 .1 .2 + 0:38z . = H z = : 1 0 0 ..

10 = Aq q10 .1 y k = B q.1 .1 + 0:2q = q .1yk = .1uk .q.1 A q = 1 .Straightforward calculations give ab3 + b4 = e.1 uk . 0:5q 1 1 .1 . d : B.1 d=9 Pole excess = deg Aq . . A q. . q =q 1 . ln a ! Problem 2.1 .1 .10 q 10 . 9 8 .9 . 1 = uk . 0:5q . 0:5q.. 0:5q . Remark 10 B q = q + 0:2 = q.d B . q + 0:2 = q.1 11 . b3 + b4 Thus and 0 0 = 1 ln ab3 + b4 + h b3 + b4 ! 1 + = 4 . CCS p.15 yk . yk + 10 .1 . deg B q = 9 = d cf. 0:5q9 q. 49. 0:5q9 System order= deg Aq = 10 B q = q + 0:2 Also . 0:5q. .1 = 1 + 0:2q. ln a ln ab3+ bb4 b3 4 where it has been used that = . 0:5q 9y k = q + 0:2uk Aq = q10 .9 + 0:2q .1 + 0:2q. 0:5y k . 9 + 0:2uk .1 A q. . h. 0:5yk + 9 = uk + 1 + 0:2uk q 10 . 10 .

. Observable canonical form: = . : : . F1 . F1z = 12 . 1 + : : : + bnuk qnyk = b0qn + b1qn.2 page 57: y 0 = 0:5 y .1 n n. y1z.n yk = b0 + b1q.1 H q = b0q + b1q qn + : : : + bn = b0 + b1q +n: : : + bn q q H q = D + Bq A n:th order system b.a2 0 1 : : : 0 = 0 0 1 1.1 X j =0 . bn 0 8 b1 9 Problem 2. y0 . 1:5qy k + 0:5yk = quk Use the z -transform to nd the output sequence when uk = 0 k 0 1 k 0 Table 2.a 0 0 8 n 9 0 C = : 1 0 : : : 0 .1 + : : : + bnq .1 . 1:5z Y . 1.17 y k + 2 .1 = 1 Z qnf = zn F . . n.1 + : : : + bnq. .Problem 2.16 FIR lter: . H q .1 = b0 + b1q. .j .1 + : : : + bn uk n.a1 1 0 : : : 0 9 8 0 1 0 : : : 0 9 0 . q2yk . D = b0 8 . y0 + 0:5Y = z U . 1:5y k + 1 + 0:5yk = uk + 1 . 1 + : : : + bn uk .= . z2 Y . . u0 f jhz.n uk = = b0uk + b1 uk . : . n yk + n = b0uk + n + b1uk + n .

Compute

y1 =u0 , 0:5y,1 + 1:5y0 = 1 , 0:5 + 0:75 = 1:25

z2 Y , 0:5 , 1:25z,1 , 1:5zY , 0:5 + 0:5Y = zU , 1 ,z2 , 1:5z + 0:5Y , 0:5z2 , 1:25z + 0:75z = zU , z 5 2, z Y z = z 20:,z1:5z 0:50:5 + z 2 , 1:z z + 0:5 U z = + 5 z = z 0:51zz, 1:5 + z , 1zz , 0:5 U z , ,0 z step U =z , 1 2 z z 2 Y z = z 0:50:5 + z , 12zz , 0:5 = z 0:50:5 + z , 12 + z ,10:5 , ,

Table 2.3 page 59 gives via inverse transformation.

z = e,k=T ; e,1=T = 0:5 T = 1 ln 2 z , e,1=T

Z ,1 z,1 z ,z0:5 = e,k,1ln2

1

1 + e.k.1 z . 12 = Z .1 ln2 Z .1 z .1 z.kln2 + 2k .1 . z Z . 12 = k . 1 h = 1 yk = 0:5e.

5 : y.25 y0 =0. 0:5y 0 = 1 + 1:5 1:25 . 4 = 21 2 4 8 13 . ln 2 = 8 + 2 + 1 = 21 2 8 y 2 = u1 + 1:5y 1 .2ln2 + 2 + e. 0:52 1 = 1 + 3 5 .1 =1 Checking the result: 1 y 2 = 0:5e. 8 y1 =1.

18: Verify that Z 2 .Problem 2.

example A1: H z = C zI . The sampled system. 1 U s = 1 s 1= 1 Y s = Gs s s3 2 1 Y z = h zzz. . z . In that case there exists a power series expansion and termwise di erentiation is allowed. P .k = X .k = . X k2 z. 1 .1 kh2 cf.k z multiply by z kz = z .+13 z . Double integrator: The rst step is to translate Gs to the corresponding pulse transfer operator H z .1 F z = 1 kh2z. .2 2 z dz 1 2 d2 z : dz2 z . 12 U 2 z 2 cf. 13 z 2 z + 1 = z 2z 13 .1 is that f is analytic for jz j 1.kz .k.+13 Table 3:3 2 Y z = H z U z 2 + 1 z 2 1 H z = Y z = h zzz.k = h f z . 1 . 1z . 8 d2 .k .k = h 2 k=0 2 k=0 2 1 X k=0 h2zz + 1 = d z.+13 2 Remark A necessary condition for convergence of f z.14 = z .k = z 2z 13 .k. Table 2.3 2z . 14 . di erentiate twice 2 multiplication by z 2 gives k2 + kz . 1 = h zz.. 1 = 2z. 12 z ..k = z .k. Use the method of page 58.2 = Pk2 + kz . 1 . 13 . 12 X . 12 z z . 12 = zzz.1 = d z = z . z uk = step U z = z .1 = h zzz. 13 1 1 2X 2 . 1 z .k. z = .1 = X k2z .k f z 0 h2 2 1 F z = 2 f z.1 dz dz z .1.

Problem 2.1.ah e . = z . 1 .e This is the same result as obtained by using Table 2. The normalized motor has the transfer function Gs = ss 1 1 + and we get .a 1 bc = e . 1 .ah Ress=. b. e.19 a. 1 s bc a e + .ah = bc 1 .30 gives sh H z = z . The transfer function of the continuous time system is Gs = s bc a + Equation 2.ah a z.a e s.ah .

esh s ss + 1 . H z = X s=si 1 Res esh . 1 1 z .

h = z . 1z .1. he z . 1 + hze.h1z.1e.h e.h . Compare Problem 2.h + .h .11. 1 = .h .h + = e . To compute the residue for s = 0 we can use the series expansion of esh . 1=s and get .h . 1 h z .121 + z . e. e. e1 . . h H z = z . e = z s+ is lead if s+ Consider the discrete time system z+b z+a .

ei!h + b .

cos !h + b + i sin !h arg ei!h + a = arg cos !h + a + i sin !h =

sin !h

**sin !h
= arctan b + cos !h , arctan a + cos !h
**

15

Problem 2.21

0:5 Simulation of the step response of the system in Problem 2. 0 dotted.2 b = .22 for 10 Time 20 lower solid.10 Output 0 0 Figure 2. .0:9 upper solid. and 1 lower dashed. Phase lead if .0:50 dash-dotted. .0:75 upper dashed.

0. Problem 2. 2.0:5.0:75. .ah xkh + 1 . xk Simulation of the system for b = .23 A state space representation of Gs is x = . .ax + b . x + : 0 . au _ y =x+u The assumption that the system is stable implies that a 0. e.22 Problem 2. 0:5 and 1 is shown in Fig. Sampling this system gives xkh + h = e. u 9 1 8 yk = b + 1 : 1 b .0:9. a ukh a y kh = xkh + ukh 16 . sin!h sin !h arctan b + cos !h arctan a + cos !h sin !h a + cos !h 0 !h sin !h b + cos !h We thus get phase lead if b a.0:4 9 8 1 9 xk + 1 = : 1 0 .2. A state space representation of the system is 8 1:1 .ah b .

e ah = q q +. a 1 .28 y k = yk . b 2a b 2a The inverse is stable independent of h: Stable inverse if ah lnb=b . Problem 2. e.. aq1 . 1 1 a or b 0 a 1 . e.ah .ah .ah The inverse is stable if = b .e where b . e. e. e.ah H q = b . 2a. 1 = 0 which has the solution The solution has the form z12 = 1 2 5 p y k = c1 Using the initial values give . e.ah . 1 + y k . 1 .The pulse transfer operator is . e. z .ah .e =a + 1 = .ah 2 Since a 0 and 1 .ah = a b. 2 y 0 = y 1 = 1 The equation has the characteristic equation z2 .1 .ah 0 then 0 b For b 0 we have the cases 2a 1 . = a 1 .

1 + p5 k 2 + c2 .

1 . p5 k 2 1 p c1 = p 5 + 1 2 5 1 p c2 = p 5 . 1 2 5 17 .

2 yk = . 0:5q + 1 yk = 2q + 1 uk The system is of order 11 = deg Aq and has the poles p 1 i 15 p12 = 4 p3:::11 = 0 multiplicity 9 and the zero z1 = . 008 qq : The second part can be obtained by sampling a rst order system with a time delay. H4 can be rewritten as : . 0:5q.30 Sampling G1 with h = =! gives.e. This implies that H3 cannot be obtained by sampling a continuous time system.The system is given by . q9 q 2 .. !0 h i. see Table 2.10 + q. 18 The system H1 has a pole on the positive real axis and can be obtained by sampling a system with the transfer function Gs = s K a + H2 has a single pole on the negative real axis and has no continuous time equivalent. This gives .1 . 2 We have two possible cases !2 i G1s = s2 + 2 !0 s + ! 2 0 0 s ii G2s = s2 + 2 ! s + ! 2 0 0 .1 + q.29 Problem 2.2q.11 uk Multiply by q 11.0:5 Problem 2. we get a pole zero concellation. :8 H4q = 2 + 09q. Compare Example 2. Sampling G2 gives H z = 0. The only way to get two poles on the negative real axis as in H3 is when the continuous time system have complex conjugate poles given by 2 s2 + 2 !0s + !0 = 0 Further we must sample such that h = =! where p ! = !0 1 .1 1+ H z = 1 +q + q + = q + 2 where = e.8.

1 1 : s .0 + . _ 1 1 0 9 8 9 8 9 8 = 0:2 h = 0:3 0 H q = C qI .1q .s 1 + . eh. .q0.1 8 1:350 0 9 = : 0:405 1:350 . ds = : ses .0 = : ses .3h q.1 . 8 0:245 9 .0 = : 0:005 .1 = e se. 1 x = : 1 0 .Problem 2. With h = 0:02 we get : : H q = q . . e. h + + h . = L.sI . : 1 + h . eh. . .3h H q = 1 . : es . + 1 .000392q . heh eh h. 1 . : .h e. 0 .1 + e . . 8 s 9 8 es 9h. = Ah.1 : .1 . e.1.32 8 The pulse transfer operator is given by where 9. .1 s . 8 0 9 Z Z e . eh. ut .0377 :9802 0:9418 Problem 2.s 9 8 eh. . h. A.1 + e = : h. 98 9 Z 8 . 1 . = 9 8 1 8 eh 0 9 0 s. 19 . 12 s .1 : s . 1eh. ds = 0 0 8 9 80 .1 =: =L . 9 h.e 3 q. 1e h .1 + eh. 1 e =: = h .31 We can rewrite Gs as 1 1 Gs = s + 1 + s + 3 Using Table 2.1 gives . e 1 .1 = : 0:050 . 8 0:105 9 . . x + : 1 . : e . 0 8 9 eh.h + 1 1 . ds = : .1 . 1eh = L. eh.

0:15 .0:39 There are two roots outside the unit circle.2 0:5 0:5 .0:5 2 . 1 .2 0:75 2 = 2=3 5=12 . 1 .1:5 1 2 = 0:9 0:19 .3 2 .0:57 .0:5 0:75 .0:5 0:75 .1 1 1 .2=3 5=12 1 = . b.2=3 . 20 .2 1 3 = .3 1 3 = .0:5 .13=20 One of the underlined elements is negative and there is thus one root outside the unit circle.0:5 2 .0:15 0:19 1 = . a.0:58 0:33 0:33 .0:15=0:19 = .0:5 . 1 .1 0:75 2 = 1:33 .0:58 1 = . The roots are 0:35 and 0:82 0:86i.1 Jury's criterion is used to test if the roots are inside the unit circle. The roots are 2:19 and 0:40 0:25i.Solutions to Chapter 3 Problem 3. The roots are 0:75 0:58i.2 2 .0:79 0:07 The roots are inside the unit circle since the underlined elements are positive.8=5 .1:5 0:9 0:9 . c.

0:4 + K = 0 K 0 The stability can be determined using root locus. The starting points are z = 0. a2 .1:7 1 3 = . a2 . 0:6 4a2 . 0:2 and 0:4. Then a + iba + ib . To nd where the roots will cross the unit circle let z = a + ib. and 0:5. 0:6a . The roots are 0:7 and 0:5 0:866i.0:5. 1:2a . 0:92 = 0 21 . 1 .0:51 0:51 2=1 0 0 0 0 The table breaks down since we can not compute 1 .0:7 1:7 .1:25 . .0:25 ..a2a .5.0:56 4:69 6 6 4:69 .Ka b2a .0:51 0:51 0:51 . 0:6a . ib Equate real and imaginary parts. ib and use a2 + b2 = 1. b2 + 0:08 = . where a2 + b2 = 1.K a . 0:4 = .1:25 . One root is outside the unit circle. . 1 . a2 + 0:08 = . There is one of the underlined elements that is zero which indicates that there is at least one root on the stability boundary. 0:6a .0:25 5 1 . b2 + 0:08 + i2ab . 0:2z .d.0:7 .K Multiply with a .1:7 1:7 .2 The characteristic equation of the closed loop system is zz . e.0:7 0:51 . The roots are . 1 5 . 0:6b = .1:25 2 = .0:56 63:70 54:92 54:92 63:70 16:47 3 = . The crossing point of the asymptotes is 0:2. 0:6 = Kb If b 6= 0 then a2 . The asymptotes have the directions =3 and . The complex conjugate roots are on the unit circle. 0:2a + ib .10:71 1 = 0:86 Problem 3.

1 The root locus for the system in Problem 3. = Kekh .0::652 0 352 . i. . a = 1. 0 Re 1 The solution is This gives K = 2a . 0:4 + K = 0 K = . 1 The regulator is Problem 3. 0:4 + K = 0 or There is a root at z = 1 when or K = 1:68 11 . 0:21 .1:30.1 is obtained when n p a12 = 0:15 0:0225 + 0:23 = 0:15 0:502 = . 0:2. The root locus may also cross the unit circle for b = 0.1 Im 0 −1 −1 Figure 3. Sampling the system Gs gives the pulse transfer operator h H q = q . 3.e. where K 0 and = 0 or h.1.1 .3 ukh = K uc kh .1.1 . y kh . 0:6 = 0:70 and .2.0:48 The closed loop system is thus stable for K 0:70 The root locus for K 0 is shown in Fig. 22 . . A root at z = .

1 = 0 The system is thus stable if K 2=h When there is a delay of one sample = h then the regulator is ukh = K ekh q and the characteristic equation is Kh + z z . The system is thus stable if K 1=h b.s s The phase function is arg Go i! = .=2 . The transfer function for the open loop system is Go s = K e. Consider the continuous system Gs in series with a time delay of seconds. w 2 and the gain is jGoi!j = K ! The system is stable if the phase lag is less than at the cross over frequency jGoi!cj = 1 !c = K The system is thus stable if The continuous time system will be stable for all values of K if = 0 and for K =2h when = h. When = 0 then regulator is ukh = Kekh and the characteristic equation of the closed loop system becomes Kh + z . !c . 1 = z2 .a. =0 K =2 = 1 =h 2h . z + Kh = 0 The constant term is the product of the roots and it will be unity when the poles are on the unit circle. This value is about 50 larger than the value obtained for the sampled system in b. 23 . .

i sin ! cos2 ! + sin2 ! + 0:25 .1 Im 0 −1 −1 Figure 3. The argument is . Problem 3. sin = cos ! :. 0. .4. 0:5 = = cos ! . 0:5 . cos ! : . In this case 1 Hoei! = ei! . 0:1 + i sin ! 5 cos ! .1 = .5cosi! ! 1 25 For ! = 0 then Ho 1 = 2 and for ! = then Ho .2=3.4 The Nyquist curve is Ho ei! for ! = 0.2 0 Re 1 The Nyquist curve for the system in Problem 3.

1:32 =3 0 .2.0:67 0 The Nyquist curve is shown in Fig.arctg cossin. 0:5.0:50 .0:57 . 0:5 ! The argument is . ! arg Ho = . The following table gives some values for the real and imaginary parts. ! ReHo ImHo 0 2 0 =6 0:97 . 3.0:80 2=3 .0:40 .=2 for ! = =3cos ! . 24 .1:16 =2 .

25 8 9 81 1 1 Wc = : . The controllability matrix is 8 9 8 9 : . . u0k In this case rankWc = 1 and the system is not reachable from u0 . = : 6 1 . 1 = : 1 .5 Consider the system 81 xk + 1 = : 1 8 yk = : 0 0 xk + 1 uk :0. xk 9 8 9 We have y1 = x21 = 0 y2 = x22 = x11 + x21 = 1 Further x11 = 1 81 09819 819 829 x2 = x1 + .4 9 W o = : C .7 The controllability matrix is 1 0.1 = : 3 . xk + : 1 . . b. = : 1 . Problem 3. the rst two colums are linearly independent. The possibility to determine x1 from y 1. This implies that Wc has rank 2 and the system is thus reachable. 1. The observability matrix is 8 C 9 8 2 . The system is not observable since det Wo = 0.2 .6 a. Wc = 4 1 det Wc = 2 and the system is reachable. + : 0 . For instance. : 0 . : 1 . = : 1 0 0:5 9 80 0 9 Wc = : 1 0:5 . From the input u0 we get the system 81 0 9 809 xk + 1 = : 0 0:5 . 9 1 . y 2 and u1 is called observability. . + : 0 .Problem 3. . . .u1 = : 1 1 . Problem 3. 81 09829 819 819 x3 = : 1 1 .

2.11 The closed loop system is H yk = Hcl quc k = 1 + cHH uc k H c a. 0 1 0 0 2 9 8 3 9 8 0 9 8 3 + u0 9 3 . : 3 + u0 . .T x2 = u0 = . = : 3 + u0 . + : u0 . = : u1 .T is not in the column space of Wc and can therefore not be reached from the origin. uk 0 0 9819 8 0 9 8 3 9 2 3 . xk + : 1 . 0Kq + K uck :5 Using the conditions in Example 3. : 1 . . With an integral controller we get Hclq = qq . With Hc = K where K 0 we get y k = q2 . q+ 26 . 0 0 0 8 not full rank not reachable. 80 1 xk + 1 = : 0 0 0 0 80 1 x1 = : 0 0 80 x2 = : 0 0 0 0 1 0 0 809 29 3 .3 u1 = 0 b. Two steps. but may be controlled.Problem 3. 9 : 1 1 1 . Problem 3. = : 1 0 0 . 1:5Kq 0:5 + K . The steady state gain is K Hcl 1 = K + 0:5 b. It is easily seen from the state space description that x3 will be 0 for all k 0. 0 0 0 0 8 9 : 0 0 0 . in general it would take 3 steps since it is a third order system. c. 1qKq0:5 + Kq = qq2 . 8 9 8 9 0 1 0 Wc = : .2 we nd that the closed loop system is stable if K 1.8 a. + : u1 .

12 0 K 0:5 The steady state gain is Hcl 1 = 1. 1 = K 0:50:5 !1 z . 2 = 0:99 rad s The sampled system will not detect this oscillation if h = k2=! .1 .1 + 1:5 The z -transform for a ramp is given in Table 2. q . + . 2 q where = exp. and we get z Uc z = z . 12 = 0:5 lim z k!1 z . 1 1 . 12 Using the pulse transfer functions from Problem 3. 2. b. To formalize the analysis we can sample the system with h = 2=! .e. The frequency ! will then have the alias ! 0 = 0. Hc z Uc z !1 z if K is chosen such that the closed loop system is stable. z lim ek = z!1 z . H q = 1 . To use the nal value theorem in Table 2. 1 z z 2 0:50z:5z K K K z .q . a.3 shows the continuous time and the sampled step responses when h = 2=! = 6:3148. The pulse transfer function is Table 2. i. z . also compare Example 3. + + k!1 The derivative is positive in steady-state and the reference signal and the output are thus diverging. + .11 and the nal value theorem gives . y k k!1 !1 .13 Gs = s2 +s0+21 + 1 :s The damping of the system is = 0:1 and the undamped natural frequency is !0 = 1.0:1h. = zlim z . 3. There is a pole zero cancellation and only the rst order exponential mode is seen at the sampling points. 27 Consider the system .1F z does not have any roots on or outside the unit circle. For the proportional controller we must then look at the derivative of the error. + :5 + K Problem 3.3 it is required that 1 . Fig. lim ek = klim uc k .13. = 1 . Problem 3. z lim e0 k = zlim z . The step response of the system will have an oscillation with the frequency p p ! = !0 1 . 1 z z z 2 1:51z:5z 0:50+ K K K z .The system is stable if 0:5 + K 1 and we get the condition and 0:5 + K .3. For the integral controller we get 2.

1 .0:54 K 12:31 The steady state gain is 0:056K Hc 1 = 0:03 + 0:056K The steady state error when the input is a unit step is 0 03 e1 = 1 .1 + 1:65 .3 25 Time 50 Problem 3. Figure 3.2 Output 1 0 0 The continuous time solid and sampled step dots responses for the system in Problem 3. 1:65 + 0:030K 0:68 + 0:026K .13. 1:65qK 0::68 + + 0026q + 0:026 +0 K :030 The characteristic equation is z2 + 0:030K . When Hr = K then the pulse transfer function for the closed loop system is 030q 0: Hc q = q2 . 0:030K This gives the stability region . Hc 1 = 0:03 +:0:056K The minimum error is about 4. 28 . 1:65z + 0:68 + 0:026K = 0 The closed loop system is stable if 0:68 + 0:026K 1 0:68 + 0:026K .14 a.

5 250 Time 500 b. The characteristic equation is then z 2 . 3.4. 29 . c. 1. and 0:05 are used on the tank system.5 and 3. 1:65z + 0:68z . If the closed loop system is stable then the steady state error is zero if the integral controller is used. 3. 1 + Kz 0:03z + 0:026 = 0 A scetch of the root locus is shown in Fig. Fig.14 is controlled with an integral controller. and 2 dash-dotted.2 1 Im 0 −1 −2 −3 Figure 3.6 show the step responses when the proportional controller with K = 0:5. 1 dashed.14 when the proportional controller is used with K = 0:5 solid. −1 Re 0 1 1 Output 0 0 Step response of the system in Problem 3. 0:025.4 −2 The root locus when the tank system in Problem 3. Figure 3. and 2 and the integral controller with K = 0:01.

1 Output 0 0 Step response of the system in Problem 3. K 0:84 16 K 0:3 = 0:053 16 K .16 The pulse transfer function for the closed loop system is H Hcz = 1 + o zz H o and the characteristic equation is z 2 + z + 0:16 + K 4z + 1 = z2 + 1 + 4K z + 0:16 + K = 0 Using the conditions in Example 3.1 + 1 + 4K 0:16 + K . Figure 3.1 . 2:5 = .14 when the integral controller is used with K = 0:01 solid. 4K This implies Since it is assumed that K 0 we get the condition 0 K 0:053 for stability. 0:025 dashed.0:432 30 . 1 . and 0:05 dash-dotted.2 give 0:16 + K 1 0:16 + K .6 1000 Time 2000 Problem 3.

. To = W 3 1 11 22 .2 . : 5 6 . = : 1 3 . Example A.17 we nd that the transformation is given by ~ Tc = Wc Wc. . = 11 : 1 .17 Using 3. = : 3 11 . = : 5 4 = 39z + z z2 . = : 0 1 . These zeros may very well be outside the unit circle.1 Wo = : 1 0 . : . 63. 3 ~ = :3 0. 2 .18 p. 0 8 9 8 ~ C = : 39 4 .1 ~ where Wc is the controllability matrix of the original system and Wc is the controllability of the transformed system. : 3 11 . 63 CCS. . 4 11 4 3 1 9 9 In the same way the tranformation to the observable form is given by .1 1 3 . new" zeros appear as described on p.1 z . 0 1 The transformation to give the controllable form is thus 81 Tc = : 0 98 9 8 9 . 9 8 9 8 : .3 p. 65 CCC Sample the harmonic oscillator. Wc = 4 11 8 The pulse transfer function of the system is H z = C zI . see a When a system with relative degree 1. : 4 . CCS 31 . . = : 5 6 . .1 ~ o. 81 09 For h = 2=!. This mapping maps the left half plane on the unit circle. 8 9 8 9 ~ ~ ~~ Wc = : .Problem 3. not controllable See e See p.2 9. 530 CCS.4 4 8 9 8 9 8 Probem 3.1. 1 .18 a i b c d e f g i iii iii ii ii iii Poles are mapped as z = esh . See Example 2. Thus the transformed system is given by 8 9 ~ . 1 0 and 9 8 z . = :1.1 8 3 9 6 .

z lim ek = zlim z .7.20 Ho = q2 +10:4q Hr = K a. E z = 1 . .1 . K = 0:5 the nal-value theorem can be used.g. + 32 . = : 1 2 0 . .1 + 0:4 K .Problem 3. 2.21 + yk = q2 0:41q:2q 0:80:5 uk .0:6 K 1 ek = uc . 0 0 0 and the origin can be reached from any initial condition x0 by the control sequence u0 = u1 = u2 = 0. y . H Htot = 1 + r Ho = q2 + 0Kq + K Hr Ho :4 The system is stable if. The system may be controllable if the matrix is such that n x0 = 0. b. 82. K 1 K . see Theorem 3. Htotz Ucz If K is chosen such that the closed loop system is stable e. z 2 + 0Kz + K z . 1 1 . 1 E z = zlim z .19 a. see p. 1 = !1 z !1 z k!1 :4 2+ :4 = zlim z 2 z 0:40z:4z K = 1:41+ K = 0:74. In this case 80 0 09 3 = : 0 0 0 . 0:4 b. 2 3 0 Since one column is zero we nd that the system is not reachable det Wc= 0. K = 0:5 !1 + + Problem 3. . Probem 3. The controllability matrix is 8 9 8 9 0 1 0 Wc = : .

1:2 0:5 + 0:8K . 3. see p. Fig.21b. :8K H Htot = 1 + r Ho = q2 . 0: q + H Htot = 1 + rHo = qq2 .1 +0 The system is stable if. 1:2q +40:5 0:8K :4q + 0:8 = Hr Ho + K 0 4q = q 3 .0:25 K 0:625 b. a.7 0 2 The root locus for Problem 3. 0:5 + 0:8K 1 0:5 + 0:8K . 82 CCS. we can determine that the closed loop system is stable if p . 1:2q 2 0:q 0+ 0:8:KK + 0:8K + :5 + 0 4 Using root locus.0:25 .1 + 0:4K .2 Im 0 −2 −2 Re Figure 3.0:25 K 16 Hr = K q 33 .6:75 K .7. 0:4K + 1:2 K 0:625 K . 1:2q0:4q5+ 0K 0:4q + 0:8 = Hr Ho + 0: + 0 = q 2 + q 0:40:4q +:2:8K:5 + 0:8K K.1 .17 + 489 .

b 9 8 p + tr 9 21 1 : `1 . This gives 8` 9 8 98 9 8 9 . = : p21. a12a21 + a12b2 . = : b1 a11b1 + a12b2 . . The solution is where = a21b2 . : .1 we nd that = 0:5 34 . a11 1 2 To check when 6= 0 we form the controllability matrix of the system 9 8 9 8 : .1 . b1 = 0:5.1 1:5 This is the same as given in Example 4. a12 a21. = : . a11 + a22 . : `1 .L = z 2 .4. 0:35z + 0:025: Using the result from Problem 4.1 . a21 = 0. For the double integrator with h = 1 and dead beat response we have a11 = a12 = a22 = b2 = 1. Problem 4. b1`1 z + a11 a22 . 0:1z . . 22 1 2 where tr = a11 + a22 and det = a11a22 .0:5 . 2 8 ` 9 1 8 a b . a12b2 + b1b2 a22 . a22b1`1 + a21b1 . 0:25 = z 2 .Solutions to Chapter 4 Problem 4. a11 b2`2 Identifying with the desired characteristic equation gives 8 b 9 8 ` 9 8 p + tr 9 b2 1 : a12b2 .0:5 . det . . b2`2 .2 In this case the desired characteristic equation is z . a22b1 a21b1 .a12b2 + a11b2 b12 . .1 The characteristic equation of the closed loop system is detzI . : p21. = : 1 .1 : `2 . Wc = b2 a21b1 + a22 b2 and we nd that = det Wc There exists a solution to the system of equations above if the system is controllable. det . a b . : 2 . and p1 = p2 = 0. 1 = . a11b2 .

and L is obtained from 8 9 8 98 9 8 9 T = : `1 .h 2 +1. +21 .h 2 = : 1.25 and 0. : 0 0 . 21 + + 21 . .21.1. h . . . 8 1 . 9 .h h = 2 21 . The dead beat controller is characterized by p1 = p2 = 0: From Problem 4.L = : 0:5 0:1 . 0 . h .2 2h e This is a nonlinear equation of the form h = f h 35 . 1+ . 1 + = h1 . 2 9T . 1.2 we have 8 = :1. h1h. . = : 0:75 .L thus has the desired eigenvalues 0. L `2 0:1 1 . . = :h. L 1. 9 8 1. . 8 = 1 . = : 0:5 0:1 . 2 or . 2 h . The matrix . where If x0 = : 1 1 .1 it follows that 8 2 T = 1 : 1 .3 For the motor in Example A. where = e. 2e . 3 + 1 . 8 9 1 1. + 1 = 2e. . 1 + 1 . 1 . h . `2 = . 9 1 = : . 1. .h .1. This function is shown in Fig.`1 . We thus want to nd h such that 1. : 0:75 . then 981+ 9 . 2 u0 = . Problem 4.h+ 1 2. . = 1 h1 . 1. 4.0:025 0:1 To check the result we form 8 1 0:1 9 8 0:75 0:1 9 8 0:25 0 9 .: . = 1 : 0:5 0 .

0 0. Using the results in Problem 4. 0029 .63 0.3. 0.ga.21 L=placefi. : .67 . 4.0:0880 . Fig.1 u0 s function 2 3 Sampling period h 4 5 of h in Problem 4. = : 9:22 .0:0125 0: 0:1595 3:11 8 98 9 8 9 The problem is easily solved using Matlab by giving the commands fi= 0.0 Output u(0) −2 0 1 Figure 4.55 0.0:1 : .16 .01..4 a.00100 .2 shows the response of the motor controlled with a deadbeat controller when h = 2:21.12.0:5900 .1 gives LT :1600 = . P=roots 1 -0. ga= 0. 8 9 Problem 4. We see that the system settles after two samples and that the constraint on the control signal is ful lled.P 36 . One way to nd h is to use the iterative scheme hk+1 = f hk Starting with h0 = 2 gives k 0 1 2 3 4 hk+1 2 2:26 2:20 2:21 2:21 with h = 2:21 we get L = : 0:49 0:51 .

: 0:03 . Wo. : 0:22 1 . and h = 2:21.2e. .21 rc=Logrd h Ac=polyrc p The chosen sampling interval is higher than recommended by the rule of thumb.3 when x0 = : 1 0 . : 1 0 .4 we nd that the continuous-time characteristic polynomial s2 + 2 !s + ! 2 corresponds to z 2 + p1z + p2 with p1 = . 5 Time 10 b. 809 8 0 9 = : C . In this case 8 C 9 8 0 19 Wo = : C . Wo.2 !h = 0:21 This has the solution 0:7 and ! 5:6. is shown in 8 9T Fig.1 8 0:22 9 8 0:78 0 9 8 . .63 0.1 = : .2 8 9 Deadbeat control of the motor when xT 0 = : 1 1 .1 Output 0 −1 0 1 5 10 Input 0 −1 0 Figure 4. 37 . 2 = .4:55 4:55 9 8 0 9 8 0:114 9 = : 0:03 . !h cos!h 1 . so the characteristic polynomial becomes s2 + 7:8s + 31:7. = : 0 . From Example 4.41:55 8 4:55 0 . 4.5 a. . = : 0:22 1 . = : 0:03 . Problem 4. In Matlab you can do rd=roots 1-0. . since !h 1:1 0:6 8 9 c. . 9 = . The closed loop system when using L = : 9:22 3:11 .0:63 p2 = e. .

Further 8 1 . and L = : 9:22 3:11 . CK = 1 ii. 1 + .k1 0 0 . KC are in the origin.3:55 3:55 9 8 yk .32 has the form xkjk = I .1 . : 0:22 0 = 0:78 . 0. 1 + Kyk: ^ x In this case we want to nd K such that i. I .28 has the form 8 9 8 .. + : 0 . 1 + . c. KC has eigenvalues in the origin. : 9 8 9 38 . 1 xk + 1jk = . + uk .1 but with T and C T instead of and . The reduced order observer 4.k 9 8 0:78 I .1 Output 0 0 5 0 Input −5 −10 −15 2 4 0 The 9 8 response and the control signal of the system in Problem 4. KC ^k .uk + Ky k: ^ x In this case we choose K such that the eigenvalues of . Using the results from Problem 4.4 8 9 T when x0 = : 1 0 .1 : yy . The rst condition implies that k2 = 1. KC = : 0 0 1 . 1 9 8 0:114 9 =: 0 1 . : y k . 1jk . 1 ^ k b. Figure 4.3 2 Time 4 We get k xk = Wo. KC ^kjk . gives 8 2:77 9 K = : 1:78 . uk . The dynamical observer 4. 0:22k1 .uk .

1 + Ky k ^ x uk = . KC . 0 407 8 9 8 9T Fig. and when ukh is zero up to t = 250 and one thereafter. KC ^k . xkh The continuous time poles of the system are . 1+: 1 .0:039412 = 0:62 The desired characteristic equation of .3:55 9 8 0:114 9 8 3:55 9 xkjk = : 0 0 .3:55 3:55 9 8 yk . 4. 1 which is the same as the observer obtained by direct calculation. xk . Ky k ^ 39 . KC is thus z2 . The observer should be twice as fast as the fastest mode of the open loop system. 1:24z + 0:38 = 0 Using the results from Problem 4. 1jk . 1 9 8 0:114 9 xkjk = : 0 ^ 1 .0:0197 and .7 The observer and the controller are described by xkjk = I . We thus choose the poles of the observer in z = e.The eigenvalues will be in the origin if k1 = 0:78=0:22 = 3:55: The observer is then 8 0 . + : 0 .4 shows the states and the estimated states when x0 = : 1 1 . xkh + : 0:0296 . y k ^ ^ Since x2 kjk = y k we get ^ 8 . uk .uk . ukh 8 9 ykh = : 0 1 . : y k . Problem 4. Problem 4.10 we get for h = 12 8 0:790 0 9 8 0:281 9 xkh + h = : 0:176 0:857 .6 From Problem 2. Introduce ^ k = xkjk .Lxkjk: ^ In the state equation both x and y have the time argument k.1 gives K = : 0::139 . 1jk .1+ : 0 .0:0129. 1 + I . uk .

KC Ky k . 1 + Kyk . 1 = I . .L k .L k . 1 = o k . 1 + I . 1 + I . LKyk = Co k + Do yk: The observer and the regulator can thus be written in the form given in the formulation of the problem.o y k . KC . . KC k . 1 + Kyk .6 The states solid and their estimates dots for the tank system in 250 Time 500 then k = I .L k .x1 and xe1 1 0.5 0 250 500 x2 and xe2 1 0.8 The constant disturbance v k can be described by the dynamical system wk + 1 = wk vk = wk The process can thus be described on the form given in 4.uk .5 0 −0.LKy k . KC .43 with w = 1. I .5 0 Figure 4. 1 + I . 40 819 xw = : 0 . 1 = I . 1 .4 Problem 4. 1 + . 1 The output of the regulator can be written uk = . . KC k .5 0 −0. KC . KC . Problem 4.

The disturbance can now be calculated from the state equation xw wk .L = : 0:179 0:142 . . Assume that L is determined to give a dead beat regulator then 8 9 : 3:21 5:57 .L . . 1. Lw wk: This gives the closed loop system xk + 1 = xk + xw wk . 1 = 1 0 xk . 1: The rst element in this vector equation gives wk . .Lw wk yk = Cxk In general it is not possible to totally eliminate the in uence of wk. L= and and 8 . 1 = xk .0:114 9 . .uk . 1 .0:142 .a. This is only possible if xw .1 . Lw = 5:356 b.Lw w1 = Hw 1w1 The in uence of w or v can be zero in steady state if Hw 1 = 0 This will be the case if 22 Lw = 1 .Lw wk = .1 xw . .Lw is the zero matrix. j element of . 'c+ ' 'c22 2 c12 1 where 'cij is the i. . xk . We will therefore only consider the situation at the output in steady state y 1 = C I . . Compared with the controller in a there is a delay in the detection of the disturbance. . The following control law can now be used ^ uk = .Lxk .Lxk . . Lw wk ^ where Lw is the same as in a. .Lxk . . If the state and v can be measured then we can use the controller uk = . In this case is the state but not the disturbance measurable.Lxk + xw .L and i is the i:th element of .uk . 41 . xk . 1 . . 1 Since wk is constant and xk is measurable it is possible to calculate wk = wk ..

The error equation is 8 xk + 1 9 8 .41 8 xk + 1 9 8 9 8 xk 9 8 . = : 0 1 . ~ ~ The characteristic equation of the system matrix for the error is z 3 +k1 . "k ^ ^ "k = y k . 2:2z 2 +1:05 . Notice the di erence in scale in the upper left curve.2 0 2 5 10 15 Output 0 −2 0 5 10 15 Estimate ve 0 5 Time 10 15 Output 0 −2 2 0 0 5 Time 10 15 The output of the system in Problem 4. The estimate of v is also shown for case c. + : 0 . C xk ^ The gain vector can now be determined such that the error goes to zero provided the augmented system is observable. b upper right and c lower left and right. b and c are used. uk + : Kw . 9 8K9 ^ ^ xw : wk + 1 . 42 .5 c.0:6433 kw = 3:3333: The controller now has to be uk = . 1:7k1 + k2 + kw z +0:7k1 +0:15 . 4. KC 9 8 xk 9 ~ ~ xw : wk + 1 . Lw wk ^ ^ where L and Lv are the same as in a.Kw C 1 .2 2 Output 0 −0. Figure 4. : wk . The solutions above have the drawback that there may be an error in the output due to the disturbance if there are small errors in the model. k2 = 0: The eigenvalues are in the origin if k1 = 2:2 k2 = . : wk .0. Fig. If only the output is measurable then the state and the disturbance can be estimated using an observer of the form 4.Lxk . 0:7kw . = : .8 for the regulators in a upper left.5 show that the output when the controllers in a.

1:55z + 0:64 = 0 Using the result in Problem 9. 1 8 0:281`c 9 + : 0:0296`c .Problem 4. = : . : x3k . 0:281`1 . 0:0261`2 .5 by augmenting the system with 8 9 x3 kh + h = x3 kh + uc kh . p 9 0 5218 0 : . = 8 0:790 . b.0::2408 . 51038pp . 0 64 This gives 8 `1 9 8 3:279 .6. We get the following system of equations to determine the state feedback vector..2:3240 + 1:55:p +p 0:64 . : `2 . The desired characteristic equation is z2 .00:0296`1 0:857 . = : . 0261 .01:0296`2 .0::0296 . .1934 . 3:028p 9 5: : : `2 .2:647 + 0:281`1 + 0:0296`2z 2 + 2:3240 .0:6770 + 0:2408`1 . 0:5218`1 . Cxkh and using the control law ukh = . 0:0296`3z + + .0:281`2 .1 give L = : 0:251 0:8962 . 8 0:281 0:0296 0 9 8 `1 9 8 2:647 . 0:0261`3 = 0 Assume that two poles are placed in the desired location and that the third pole is in p. `3 :617 + :617 43 . 1:55 . uc k 1 The characteristic equation is z 3 + .0 0261 `3 0:6770 .0:281`3 9 8 xk 9 : 0:176 . 0:0035`2 . The state equation for the tank system when h = 12 was given in the solution to Problem 4.0::0035 . An integrator can be incorporated as shown in Section 4.9 a.Lxkh . `3 x3kh + `cuck The closed loop system is then 8 xk + 1 9 : x3k + 1 .0:0296`3 .

5 dashed.11 CCS. Fig.1 Output 0 0 Figure 4. but it is a feedforward term from the reference signal.7 250 Time 500 c. 250 Time 500 1 Output 0 0 The step response for the tank process when p = 0 and when `c = 0 solid.7 shows the in uence of `c when p = 0. 3 dashed and 6 dash-dotted.6 The stepresponse for the tank process when p = 0 solid and 0. see Fig. 4. and when `c = 0. Fig. 4. 44 . 4. The parameter `c will not in uence the characteristic equation.6 shows the step response for some values of p. Figure 4.

0::1545 .0::0245 . : kw1 . 9 xw = : h 0 . we would totally eliminate v since v and u are in uencing the system in the same way. v k 1 h h where v k is a sinusoidal.0 1545 kw2 45 8 9 .35 where 8 h2=2 0 9 8 .Problem 4. k1 h h2=2 ~0 9 8 xk 9 .Lxkh . ~ 8 1 . KC 9 8 xk 9 ~ ~ xw : wk + 1 .kw1 0 w2 0 Let the desired characteristic equation of the error be z . . : wk . 81 xkh + h = : 0 9 8 9 8 9 i. Lw wkh where 9 8 L = : h12 23h .2:9022 1 0:5 0 k2 = 6 2 ..e. The error equation is then 8 xk + 1 9 8 . wkh The augmented system 9. sin ! h 9 wkh + h = : sin !oh cos ! o . and w = : . uk + : h2=2 .4 + 3:9022 9 . a dead beat controller for the states. Compare the discussion in the solution of Problem 4.10 The process is h . 0 4756 .1:1 .. Since xkh and kh cannot be measured we use the observer of the structure 9.1 . = 0 1 0 0 . .Kw C 1 . ukh :0 0 0 where = cos !o h and = sin !o h.e.k : wk .k2 1 h 0 ~ = . 5:8044 3 :9022 9022 : 2. Assume rst that the control law is ukh = .3:19022 . it can be described by 8 cos ! h . I. = : . + : 0 . Further if Lw = : 1 0 .33 is now 8 9 8 2 9 2 8 xkh + h 9 1 h h h=2 0 8 xkh 9 h h=2 0 : wkh + h . 4 = 0 If h = 1 and !o = 0:1 then for = 0:5 we get the following system of equations 8 1 0 0 0 9 8 k1 9 8 .8. ~ : . . : .4 4. xkh + : h2=2 . wkh oh 8 o9 v k = : 1 0 . : wkh .

` . Figure 4.L = : 1 1 0:72 . This gives the condition detI . or . It is seen that the controller is able to eliminate the disturbance. and Lw = 8 9 : 1 0 .L = 2 + ..e. 0:7`1 + `2 = 2 I. 8 9 46 .8 shows the states of the double integrator and their estimates when vt = sin!o t. + . 8 k = 1:9022 1 Problem 4.. 9.x1 and xe1 2 0 −2 0 10 20 30 x2 and xe2 2 0 −2 0 10 Time 20 30 The states of the double integrator solid and their estimates dots 8 9 when vt = sin!o t.11 has all eigenvalues in the origin. . We have to determine the feedback vector L such that a.1:6 + `1 + 0:63 . 0:7`1 + `2 = 0 L = : 1:6 0:49 . The controller is de ned by L = : 1 1:5 . 8 0:9 .8 which has the solution k2 = 1:1018 :kw1 = 0::2288 kw2 = 0 1877 Fig.1:6 + `1 = 0 0:63 .` 9 .

= m To get unit steady state gain we choose m = 1 b.l2 3 2 1 Deadbeat 1 −1 2 3 l1 Figure 4. if `1 = 0. + .1 . Assume that the deadbeat control in a.1 .1:6 + `1 This gives . 8 9 47 .L.0:7`1 + `2 0:37 .e. but the system will remain stable if x2 is disconnected if `2 = 0. 0:7`1 + `2 .1:7`1 + `2 .1:6 + `1 0:63 . + .11. 0:67`1 + `2 .2 0:63 . 4. 2 + 4 `2 2 + ` + 4` .9.12 The stationary gain of the closed loop system is given by stationary gain of the m C I .L = det : . . one is L = : 2:25 0 . is used.9 The stability area for L in Problem 4.3:23 0:37`1 + `2 . The deadbeat requirement implies that the characteristic equation of the closed loop system is 8 9 `2 . The closed loop characteristic equation is stable if See Example 3. 2 = detI .. Problem 4. The closed loop system will be unstable if the feedback from x1 is disconnected i.1 + . 2:25 2 = 1 2 There are in nitely many solutions. a. 1 . 0:7`1 + `2 1 0:63 . 0:5 . 0:25 + `1 4`1 .0:03 The stability area is shown in Fig.

xk = : 2 8 . Wc = 4 9 det Wc = 0 implies that the system is not reachable and arbitrary states 8 9T cannot be reached from the origin.b. we get 2 = 2:25u0 + u1 8 = 9u0 + 4u1 8 9T 8 9T u0 = 0 u1 = 2 c. and x0 = : 0 0 . 1 . 2k1 = 0:04 which has the solution 8 1:85 9 K = : 7:48 . 2:25 = . The controllability matrix is 9 8 9 8 : . 0:4 + 0:04 = 0 8 + k . u1 With x2 = : 2 8 . However. x2 = 2x0 + . 2:25 + 0:5k2 . 2. u0 + : 4 .u0 + . 48 . 2 = 2 + k1 . 0:25 . The observer should have the characteristic equation . . 0:22 = 2 .0:5 9 detI . 2k1 Identifying coe cients give the system One solution is k1 . is in the column space of Wc and the point can thus be reached. + KC = det : k 1. . x0 + : 9 . = : 1 2:25 .0:4 0:5k2 .u1 8 0:5625 1:125 9 8 2:25 9 819 = : 2:25 4:5 .

1 we must know the pulse-transfer function B z =Az and the desired closed-loop characteristic polynomial Acl z .0:4z2 + 0:42z . To use Algorithm 5.2 a. a. we know at least that z + must be a factor in Acl z . 0:5. 1:05z + 0:275 The next step of the algorithm gives Q2 = z . 0:35: This gives Q1 = z . 0:55=0:1775 A4 = 0 The greatest common factor of A and B is thus z . An where A0 = A.0:4 A3 = 0:1775z . Since the desired pulse-transfer function from uc to y is Hm z = 1 + =z + . 2:6z3 + 2:25z 2 . You easily see that. with A and Acl being rst order polynomials and B a scalar. 0:8z + 0:1 A1 = z3 . Problem 5.1 Euclid's algorithm de nes a sequence of polynomials A0. : : :. For the polynomials in the problem we get A0 = z4 . 0:95=. 2z2 + 1:45z . A1 = B and Ai = Ai+1 Qi+1 + Ai+2 If the algorithm terminates with An+1 then the greatest common factor is An . 0:11 = . Step 1. 0:08875 = 0:1775z . you can solve the equation for the closed loop system using scalars Rz = r0 and S z = s0 : AzRz + BzS z = Acl z z + a r0 + 1 s0 = z + Identi cation of coe cients gives the following equation system: r0 = 1 ar0 + s0 = 49 with the solution r0 = 1 and s0 = . 0:5 Finally Q3 = z . . 0:6 A2 = .Solutions to Chapter 5 Problem 5.0:4z2 .

1 + 1 + . a static proportional controller. The resulting control law becomes Rquk = T quck . Rquk = T quck . 1 . . S q y k uk = 1 + uc k . a + . Solution with higher order observer: The solution above is not the we must increase the order of R by one. Ao = 1 and Ac = z + . S qy k . decide to have another closed loop pole in z = . a Step 2. + . In this case. Splitting Acl into factors Ao = z + and Ac = z + gives 1 T z = t0 Aoz = Ac1 Ao z = 1 + z + B The resulting control law becomes uk = . there are no observer dynamics. The controller thus is a dynamical system.e. and the observer dynamics are cancelled in the pulse-transfer function from uc to y . This gives and the equation system becomes only one solving the original problem. Step 1. Thus. 1 c c . We can. a + . a uk . . This could have been avoided by choosing deg S = 1. the static gain from uc to y is set to 1 Hm 1 = 1. 50 . since deg Ao = 0. ay k i..Step 2. a :s0 = . To solve the equation for the closed loop characteristic polynomial z + ar0z + r1 + 1 s0 = z + z + 8 r0 = 1 ar + r1 = + :ar0 + s0 = 1 8r = 1 0 r1 = + . ayk . In this case there is a delay of one sample from the measurements y to the control signal u. u k + u k . for example. though. . Choose 1 T z = t0 Ao z = Ac1 Aoz = 1 + B With this choice of T . say. Factor Acl z as Ac z Aoz where deg Ao deg R = 0.

25 Time process with y0 50 b. but not in the response to a set point change. i.e. the process zero is cancelled by the controller.3 or through the following discussion. The closed loop characteristic polynomial is given by AR + BS . and that the design parameters are = .. B + must be a part of the R polynomial as 51 . Cancellation of process poles and zeros is handled by Algorithm 5. Problem 5. and uc k is 0 for k 25 and . and the crosses to the rst order controller. The desired closed loop pulse-transfer function is Hmz = z2 . so z + 0:7 is not a factor of Bm . The dots corresponds to the zero order controller. and z + z + in the second one.1 Output 0 −1 0 Figure 5. In both cases we get the same closed loop pulse-transfer function from uc to y since the observer polynomial is cancelled by T z : 0 Hmz = AzRBzzTBzzS z = tABzz = 1 + + z+ c Fig. By choosing Bm = 0:2z we make the steady state gain Hm 1 = 1.Bm zz+ 0:7 1:5 In this case.0:7 and = .1 for k = 25.1 shows the response when the two di erent controllers are used. = 1. You can see the e ect of the observer polynomial when regulating a nonzero initial state. the process numerator is factored as B = B + B . z + in the rst solution. 5. and the pole excess in Hm equals the pole excess in H . B + = z + 0:7 and B. where B + is the part of the numerator which should be cancelled by the controller. . It is assumed in the simulations that a = .0:99.1 The output of the = 1. First.3 a.0:5. i.e.

1:5z 0:7:7 1:5 +0 in order to get Hm 1 = 1. Letting R = r0z + r1 S = s0 z + s1 the identity then gives the system of equations 8r0 = 1 8 r0 = 1 r1 = 0:0875 . we must increase the order of the controller by one and. the left hand side is of second order. the causality condition deg S deg R = 1 A leads us to set S z = s0 z + s1 . will not su ce in this case since it would only give 2 parameters r0 and s0 to select the 3 parameters in the second order polynomial z 2 .1:8r0 + s0 = . 1:8z + 0:81 r0 + s0 z + s1 = z 2 .0:11 : 0:81r0 + s1 = 0:7 Thus. To obtain the desired + .well as Acl . 81r : s1 = 0 52 .e.0:7uk .1:8r0 + r1 + s0 = . B:5z + 0:7 = z 2 . The closed loop characteristic polynomial is now given by the identity z 2 . 1:5z + 0:7.T 02 Hm z = ARBTBS = B + z2B. and so must cl be. 1:8r0:+ 0:17+00+7s1 = 0:7 s1 = . i. Now.1:5 2125 s 1 :s0 = 0:0:1012 :0:81r0 . 0:3y k + 0:11yk . since we have 3 indeterminates r0. With this choice of R. 1:5z + 0:7 8 8 r0 = 1 r0 = 1 s0 = 0:3 . 1:5z 2 + 0:7z . 1::5zz + 0:7 + 1 we must select T z = 0:2z The controller is now uk = . add an observer pole which is placed at the origin. 1: b. s0 and s1 and 3 coe cients to set: z 2 . z S z = B + z Acl z z 2 . consequently. 1 + 0:2uck .Bm zz+ 0:7 = z21. we can solve the Diophantine equation above. 1:8z + 0:81Rz + S z = Acl z If Rz is a constant r0. Thus. 1:8z + 0:81Rz + z + 0:7S z = Acl z The simplest choice. Ao = z and Acl = z 3 . we have Rz = B + z Rz = z + 0:7 and S z = 0:3z . This gives the Diophantine equation AzB+ z Rz + B+ z B. In this case we do not want to cancel the process zero. a zero order controller. so 0:2 z + :7 Hm z = z 2 . 0:11.1:5 :s1 = .

53 .1 0 −0.1 0 20 Time 40 Control 0. = z . The ringing in the control signal in Case a is due to the cancellation of the process zero on the negative real axis.0:0875uk .3. 1 S = z + a1 + The controller contains an integrator. Using the controller S uk = Rq uck . 0:2125y k + 0:1012y k . Further T = t0 Ao = 0::2 z .2 shows the output and the control signal for the controllers in Case a and Case b.10 gives one solution to the problem R = BAm . yk q gives the closed loop system BS = Bm AR + BS Am Section 5.1 0 20 Time 40 Figure 5.1 0 −0. Further the pole of the process is cancelled. 1 + 0:1176uck .2 The output and the control signal for the controllers in Case a left and Case b right in Problem 5. Bm S = ABm With the given system and model we get R = 1z + . Problem 5.4 a. 5.2 0 0. 1 c. The controller is thus 17 uk = . Case a should probably be avoided because of the ringing in the control signal.2 0 20 40 Control 0. Fig. 1 .Output 1 Output 0 20 40 1 0 0.

Am . The characteristic polynomial of the closed loop system is AR + BS = z + z + a The closed loop system will contain an unstable mode if jaj controller can be written S = A Bm R B Am . we can solve the Diophantine 54 . Problem 5. i. The rst constraint is the same as for the polynomial design discussed in Chapter 5. With this choice of R. Bm 1. ii.3 or through a discussion like the one below.2 gave R=1 S = . By forcing the factor z .33 gives the pulse transfer operator from uc and v to y : y k = Bm uc k + ARBRBS vk A + m The design in Problem 5. b. By inspection of the transfer function from v to y we see that we must make R1 = 0 in order to remove the steady state error after a load disturbance step. Bm must contain the zeros of B that are outside the unit circle. Equation 5. and so must Acl be. Now. 1 into Rz we thus have obtained integral action in the controller. the left hand side is of second order.5 a. With Rz = z .b. the causality condition deg S deg R = 1 leads us to set S z = s0 z + s1 .a We thus get BR = 1 AR + BS z + If v k is a step there will thus be a steady state error 1=1 + in the output. 1Rz + Bz S z = Acl z z + az . The From this we can conclude that in order to get a stable closed loop we must ful ll the following constraints. 1Rz the closed loop characteristic equation becomes Azz . The design problem is solved by using the general Algorithm 5. 1Rz + 1 S z = Acl z If Rz is a constant r0. Bm must contain the poles of the process that are outside the unit circle.

It is assumed that a = . 1 r0 + s0 z + s1 = z + z + 8r = 1 r0 = 1 0 s0 = + .3 2 Output 1 0 0 Figure 5.ar0 + s1 = 8 To obtain the desired T Hmz = AzRBzz+ BzzS z = z + Tzz + = 1 + z+ we must select The controller is now T z = 1 + z + y k . 25 Time 50 equation above. and the design parameters are chosen as = .0:99. s0 and s1 and 3 coe cients to set: z + az . 1 . 1r0 + s0 = + : s1 = + a : .0:7 and = . a + 1yk . since we have 3 indeterminates r0.3 shows the controllers in Problem 5. 5.5 when the controller does not contain an integrator dots and when an integrator is introduced crosses. 1 uk = uk . 1 Fig. + .2 and the controller with an integrator. The reference value is zero and there is an initial value of the state in the process. At t = 25 a constant load disturbance is introduced.0:5. 55 .3 The output of the process in Problem 5. a + 1 a . a + + 1 + uc k + 1 + uc k .

A R BR B This gives 0 0 Bm Ao Bm Ao R R ! Hcl = 0 + Ao Am A = Ao Am A B 1 1 + BR . H B0 Ao 0 B. The pulse transfer operator of the closed loop system is T=R B0 Hcl = A0R +TB 0S = A0=B0 + S=R The design gives 0 T = Bm Ao and AR + BS = A0Am B+ or S = B + Am Ao .It is assumed that the design is based on the model H = B=A while the true model is H 0 = B 0 =A0 . 1 1 1+ R + Am H 0 .1:4 a0 0:6 56 . 1 1 Ao B . B B.41 gives Hcl = z + a10 + .6 Problem 5.a T =1+ Assume that the true process is 1 z + a0 Equation 5. aj 1 With the numerical values in the problem formulation we get . H Problem 5. H Ao Am H 0 .2 gives R=1 S = . R + H 0 . a + The closed loop system is stable if ja0 + . The design in Problem 5.7 a. 1 R m ! = Hm ! = BA m RB.

The right hand side is also shown. 0:5 2:5 1 z . The right hand side of the inequality is also shown solid.6 shows that the control signal is given by 5. H 0z H z Hff z = 1 +z + z + a 1 + a Hfb . Further the desired model is stable. The control signal is now obtained by studying the step response of Hm =H . i.e. 0:9 .52 uk = Hmq uck = 1 +q + q + a uc k Hq We may assume that both the process and the model have a continuous time correspondence. Equation 5. This implies that a and are less than zero.40 gives the inequality z Hz H z .4 1 Frequency 2 3 The left hand side of the inequality in Problem 5.6 dotted.7 when z = ei! . 0 ! for a0 = .0:955 dashed. z + a0 z .10 5 0 0 Figure 5.4 shows for z = ei! the left hand side of the inequality for di erent values of a0 . which is a stable rst order system. . j j 1. 57 Problem 5. The magnitude at the rst step can be determined either through the initial value theorem or by using series expansion and the value is 1 + . The largest value is then either at the rst step or the nal value. If j j jaj then the closed loop system is faster than the open loop system and the control signal is largest at the rst step.8 .0:9 dash-dotted and 0. m = z . b. 0::5 2:5 z. Section 5. The nal value is 1 + a. If the desired response is slower than the open loop system then the nal value is the largest one.0 9 or 1 z . 0:9 Fig. 5.

0:6 Identifying with s2 + 2 !s + ! 2 gives ! = 0:1. 1:6q +0:65 and B = 0:4q +0:3. 2!h = . 1:6q + 0:65. 58 . Acl will at least contain Ac = q 2 . Using Example 2.1 zero in Bm causes the delay from the command signal to be one time unit more than the delay of the process. other factors may be added later on. 0:7q + 0:25 cancelled process zero. A+ = 1. = A = q 2 . The rule of thumb on p.1 = 5 with Acl = A+ B + Am Acl and Acl = Ac Ao . = 0:4 achives cancellation of the process zero. !h cos 1 . 1 = = 22+2+0+0. B + = q + 0:75 and B . Bm = B. Data: The process is given by A = q 2 . but no cancellation of process poles. Bm Bm = 0:55=0:4 = 1:375 Degree condition: Remark: The fact that we cancel one zero and do not introduce any other Model following condition: deg Acl = 2 deg A + deg Am + deg Rd + deg Sd .0 2.Problem 5.14 a.3. A. The desired response to command signals is assumed to be Hm = Bm =Am = Bm =Ac = 2 .2e. Thus an appropriate sampling interval is h = 1.2 !h = 0:5 The poles are in 0:66 0:25i.6 b.15 deg Am .1:32 a2 = e. This solution demonstrates how to use Algorithm 5. deg B 2. deg Bm deg A . 130 gives !h = 0:1 . Hm 1 = 1. Step 1. 0:7q + 0:25. 0:55=q Pole excess condition: where p Problem 5. Rd = Sd = 1 since no given factors are forced into the controller.16 we get sampled data characteristic equation z 2 + a1 z + a 2 = 0 a1 = .

Step 2. Using the degree condition above we may conclude that

deg Ao = deg Acl , deg A+ , deg B + , deg Am , deg Ac = = 5,0,1,2,2 = 0 The Diophantine equation to solve thus becomes A, Rd R + B, SdS = Acl 2 , 1:6q + 0:65R + 0:4S = q 2 , 0:7q + 0:25 q Since this is of second order, R must be a constant, r0 , say. In order to solve the identity we must have two more parameters, so we let S = s0 q + s1 : q 2 , 1:6q + 0:65r0 + s0 q + s1 0:4 = q 2 , 0:7q + 0:25 This gives the system of equations

r0 = 1 ,1:6r0 + 0:4s0 = ,0:7 0:65r0 + 0:4s1 = 0:25

r0 = 1 s0 = 2:25 s 1 = ,1

Step 3. The controller polynomials are now given by 5.45:

R = AmB + Rd R = Amq + 0:75 S = AmA+ Sd S = Am 2:25q , 1 T = Bm A+ Acl = 1:375 Ac Since, in this case, Am = Ac , this factor can and should of course be cancelled in all controller polynomials, giving

R = q + 0:75 S = 2:25q , 1 T = 1:375

The corresponding degree of the closed-loop polynomial AR + BS will thus be 3 instead of 5. In this case we want to have an integrator in the controller, i.e., Rd = q , 1. This will increase the degree of the closed loop by one compared to Problem 5.15 see 5.42, which is done by having Ao = q + ao , say. This gives the Diophantine equation A,Rd R + B, SdS = Acl q 2 , 1:6q + 0:65q , 1R + 0:4S = q 2 , 0:7q + 0:25q + ao R must still be a constant which as usual will be 1 and S must be of second order: q 2 , 1:6q + 0:65q , 1 + s0 q 2 + s1 q + s2 0:4 = q 2 , 0:7q + 0:25q + a0 59

Problem 5.16

This gives

and

S q = 2:5 a0 + 1:9q2 , 0:7a0 + 2q + 0:25a0 + 0:65 Using a0 = ,0:25, 5.45 gives after cancelling the common factor Am :

,2:6 + 0:4s0 = a0 , 0:7 2:25 + 0:4s1 = ,0:7a0 + 0:25 ,0:65 + 0:4s2 = 0:25a0

R = B + q , 1 = q 2 , 0:25q , 0:75 S = 4:125q2 , 4:5625q + 1:46875 T = Bm Ao = 1:375q , 0:34375

Problem 5.17

The minimum degree solution has deg A0 = 1 and gives a unique solution to the Diophantine equation. Let us instead use deg A0 = 2 and deg S = deg R , 1. This gives the equation z + 1z + 2z 2 + r1z + r2 + z s1 z + s0 = z 2 z 2 with the solution

**R0 = z 2 , 3z S0 = 7z + 6 The controller is causal. Using Theorem 5.1 we also have the solutions R = R0 + Qz S = S0 , Qz , 1z , 2
**

where Q is an arbitrary polynomial. Choose for instance Q = ,1. This gives

R = z 2 , 3z , z = z 2 , 4z S = 7z + 6 + z2 , 3z + 2 = z2 + 4z + 8

This is also a causal controller. The closed loop systems when using R0; S0, T0 = S0 and R; S , T = S respectively are

BS0 = z7z + 6 AR0 + BS0 z4 BS = zz2 + 4z + 8 AR + BS z4 The number of zeros are di erent.

60

Solutions to Chapter 6

Problem 6.1

In the rst case it is assumed that we have a control structure as in Fig. 6.1. There are three subsystems each with the transfer function K Gi s = s + iK i and the total transfer function from uc to y is K2 G = s +G1G2 G3 G = ss + K s + KK1s +K3 + K K K K K4G1 G2 3 K3 1 2 1 2 3 4 If either of the gains Ki is increased su ciently much the closed system will become unstable. Fig. 6.2 shows the response when uc is an impulse and when K1 = K2 = K3 = 1 and K4 = 0:1, 0.25, and 0.75. A disturbance in the process will propagate in the direction of the ow. In the case of control in the direction opposite to the ow each of the subprocesses

Raw material flow u c x2

Σ

Σ

1 s

K1

Σ

1 s

K2

Σ

x3 1 s K3 1 s

Final product x 4 =y

−1

−1

−1

− K4

Figure 6.1

Block diagram for the control in the direction of the ow in Problem 6.1.

1

Output

0 0

Figure 6.2

K4 = 0:1 solid,

Impulse response for the control in the direction of the ow when 0.25 dashed, and 0.75 dash-dotted.

25 Time

50

61

6. 6. c. Problem 6. b.3 −1 −1 Problem 6. An intentional nonlinearity is introduced in the selector. 62 . A disturbance in uc will now only in uence the rst subprocess and will not propagate along with the ow. Cascade control loops are found for the cooling media ow and for the output product ow.3. Notice the order of the states.3 in CCS contains several examples of couplings of simple control loops. Feedforward is used for the level control loop where the input ow is used as a measurable disturbance.1. The square root device is thus used to remove the nonlinearity of the measurement device. Block diagram for the control in the direction opposite of the ow in has a transfer function of the type Gi .x 4 =y 1 s uc x3 1 s K3 x2 K2 1 s K4 Σ Σ 1 s Σ 1 s K1 −1 Figure 6. The system will remain stable for all positive values of Ki . The system is then represented with the block diagram in Fig. The input ow is also used as feedforward for the cooling of the jacket. A disturbance will now propagate in the direction opposite the ow. The ow is probably measured using di erential pressure which is proportional to the square of the ow. Either the temperature or the pressure is used to control the coolant ow depending on the status of the process.2 Fig. Nonlinear elements are used in the ow control loops of the product output and the coolant ow. The reader is strongly recommended to compare with the case where the disturbance appears at the nal product storage instead. a.

3:63. : : : For !0 = 2 and !s = 2=0:2 = 10 we get the angular frequencies 2 k 10 = 2. it is 6= 0 in the two . see 7. . The sampled signal has the Fourier transform. 0 signal sin !0 t is sampled with the sample interval h its Fourier transform will be 6= 0 in the points ! + k!s .Solutions to Chapter 7 Problem 7. 13:63.. 12. 63 . 8. More precisely.e. Thus. We want to eliminate the disturbance with the known and xed frequency f2 5f1. The Fourier transform of sin !0 t has its support i. Problem 7. : : : ! = 20 gives rise to .f1 . 6:37. the+ ! where! . 18. 22. 16:37. 8. 20 k 10 3:63. 16:37. All frequencies in the interval . 6:37.3 1 1 X F ! + k! Fs = h s k=.1 !s = 2h where F ! is the Fourier transform of the time continuous signal. set . 12. k = 0.2 We have the following speci cations on the choice of sampling period and presampling lter: 1. 2. 13:63.1 Which frequencies will the signal f t = a1 sin 2t + a2 sin 20t give rise to when sampled with h = 0:2? Since sampling is a linear operation we consider each component of f t separately. ! . : : : The output of the sampler is composed of the frequencies . f1 should be possible to reproduce from the samples of the continuous time signal. it equals i ! 0 . 2. 1. 2. if the points !0 . : : : .

So to achieve 20 dB in log f1 = 0:699 decades takes n 2. the disturbance does not mix with the data signal. for the disturbance f2 not to fold on the data signal fs =2 .5 and 7. If fs 6f1 . Filter out the disturbance using an antialiasing lter. 64 .1 and Fig.6 The magnitude of the spectrum of the sampled signal can be obtained by folding the spectrum of the time continuous signal around the angular frequency !N = !s=2 = =h.1 0 10 a) h=2π/10 0 10 20 30 b) h=2π/20 0 c) h=2π /50 10 20 30 Figure 7. 7. Suppose the disturbance should be attenuated 20 dB without e ecting the datasignal. It can instead be removed using digital lters.5. Problems 7. f1 f2 . fs =2 fs f2 + f1 = 6f1 Two cases: 1. f 2. See Fig. 0 10 25 50 The sampling theorem states that the rst speci cation will be satis ed if and only if the sample frequency fs is chosen such that fs 2f1 Moreover.1 Folding for di erent frequencies in Problem 7. A n:th order lter gives maximally n 20 2 dB decade. 7.2. Then sample with fs 2f1 .

The eye acts like a low pass lter. The picture will not move if !r = n !s . if !s 2!r according to the sampling theorem. − ωs − ω r 0 ω r ωs 4 3 1 2 1 4 2 Figure 7.3. The wheel then appears to rotate three times slower and in the wrong direction.6.7 when !s 2!r . 7. Aliasing will give a frequency ! = 1=3 !r.7 The rotation frequency of the wheel !r = 2r. Compare the stroboscope. A correct picture will be seen. 65 . Problem 7. If !s = !r the wheel will appear to stand still. The wheel will appear to rotate with a frequency lower than r if !s 2!r . For instance let !s = 4=3 !r .2 Folding for di erent frequencies in Problem 7.− 100 100 a) ω s =120 − 120 − 100 − 60 − 20 20 60 100 120 b) ω s =240 − 140 − 100 100 140 Figure 7. for integer values n. The frequency of the camera shutter !s = 2=h. See Fig.3 t 3 Folding in Problem 7.

We thus get the frequencies f1 = 0 f2 = !0 in the sampled signal.Problem 7.9 The signal is 1 ut = sin4!0t cos2!0t = 2 sin6!0t + sin2!0t Sampling the signal with !s = 2h = 6!0 gives the Nyquist frequency !N = 3!0. Sampling the signal ut gives the alias of sin6!0 t in ! = 0. 66 .

1a.4. 8.Solutions to Chapter 8 Problem 8. When the backward di erence is used then 1 z = 1 . backward di erence 8.1 The three transformations Euler's method forward di erence 8. Problem 8. i!h=2 Now jzj = 1 arg z = 2 arctan !h The imaginary axis is thus transformed into the unit circle in the z -plane. b. !h This represents a circle with radius 0. Tustin's approximation. 8. For Euler's method we have z = sh + 1 This implies that the stability boundary in the sh-plane the imaginary axis is translated one unit to the right.6 have di erent stability properties. sh For s = i! we get 1 1 .2 a Gs = s + a . Finally for Tustin's approximation with s = i! z = 1 + i!h=2 1 . This can be seen by nding how the left half s-plane is transformed into the z -plane. a 0 Forward differences Figure 8. If a transfer function is stable in the s-plane it will be translated into a stable discrete time system if using the backward di erence or Tustin's approximation. see Fig.1b.5 and going through the points 0 and 1. Backward differences Tustin 67 .5 and Tustin's approximation 8. see Fig. Backward di erence and c. Eulers method.1 Transformation of the left half s-plane when using a.

68 Gk s = 4 s + 1 s+2 . be poor even if the di erence equation is stable.3 The lead network should be approximated using di erent methods. 1 = 1 + ah=2z1ah=2 2 . Using Euler's method we get ah a H z = z . 1 + ah This corresponds to the di erence equation ykh + h + ah . however. 1 + ah= +a hz+1 z+1 ah=2 = 1 + ah=2 ah=2 . At ! = 1:6 rad s it has the argument 19 and the gain 2. 1j 1 0 h 2=a: The approximation may. b. 1=h + a = z . 1ykh = ahukh: The di erence equation is stable if jah . Using Tustin's approximation with prewarping gives z+1 a a= = 1 + a= H z = z . 1 z + tanah=2 + 1 Problem 8. For Tustin's approximation we get or a z + H z = 2 z . c. 1 z + ah=2 + 1 The pole of the discrete time system will vary from 1 to .a.1 when h vary from 0 to in nity. The discrete time approximation is always stable when a 0.95. 1 a= . 1 z+ +a z+1 a= + 1 where Thus a = tanah=2 tanah=2 H z = 1 + tanah=2 z+1 tanah=2 .

Tustin's approximation 2z. 1=h + 1 = 4 zz. 0:596 +2 z+1 where Within two decimals this is the same as in c. 00::667 h . 11+ 2h = 4 zz. 2= . e. 11 = 3:333 zz. 1 . a sin!h =K 1 + b2 + 2b cos!h b . 4 2 1 . 80 z . e z . Euler's method gives . 5 b. 0:775 HTW z = 4 z .0:0778 :6 2z. e.a.1 +2 hz+1 d. 1= = 3:596 z .2h .1 +1 z+1 z = 4 z 1 + 1= .. Zero order hold sampling gives ! = tan! 1h=2 7:893 1 . 00::75 z .2h .e 2 z . Tustin's approximation with prewarping z .i!h + b H ei!h = K ei!h + a = K e i!h + ae.i!h + b e +b e + b 1 + ab + a + b cos!h + ib . + HE z = 4 z . 1=h + 2 h .1 h z + 1 + 1 z 1 + h=2 .2h = 4 z . 607 All ve approximations have all the form H z = K z + a z +b The gain and the phase at ! = 1:6 are obtained from i!h i!h e. 1 .2h 0 1 HZOH z = 4 .h e =2 = 4 z . 1 z1 + 2= . 1 . c. sin arg H ei!h = arctan 1 + ab + aa + b!h !h cos jH ei!hj = K s 1 + a2 + 2a cos!h 1 + b2 + 2b cos!h The di erent approximations give at ! = 1:6 rad s. 1.. 1 . h = 3:6 zz. 69 . Backward di erences 1 + .e . 1=zh + 2 = 4 zz1 + 2h. 0::803 z. 1=zh + 1 HB z = 4 z . h=2 HT z = 4 = 4 z 1 + h .

0:15. 8.2 shows the Bode diagrams for the continuous time system and for the Euler.2 but with h = 0:05. Tustin dash-dotted. The shorter sampling period gives a better approximation. Euler's method dashed. backward and Tustin approximations. A time delay of h=2 seconds gives at the crossover frequency a decrease of ! ch ' = !c h=2 rad = 180 c h = 0!035 = 5 . 8. 0:5: 70 jH ei! j arg H ei! 2:97 24 2:92 16 2:96 19 2:96 19 3:25 22 Problem 8.2 10 −1 The Bode diagrams for the lter in Example 8.3 is the same as Fig. 15 . backward di erence dotted. 15 2 : This gives !c h = 0:17 . 8. Further we will allow a decrease of the phase margin of 5 . It is assumed that the sample and hold circuit can be approximated by a delay of h=2 seconds.4 . This approximately corresponds to a decrease of the damping by 0:05 . 10 Frequency (rad/s) 0 10 1 10 2 Euler Backward Tustin Tustin with prewarping Zero order hold Fig.3 when h = 0:25 continuous time lter full. 0:52 or approximately !c h 0:15 . Fig.10 1 Gain 10 −2 10 30 0 10 −1 10 0 10 1 10 2 Phase (deg) 20 10 0 −2 10 Figure 8.

T =N HD z = T z .2 but when h = 0:05.3 The same as Fig.10 1 Gain 10 −2 10 30 0 10 −1 10 0 10 1 10 2 Phase (deg) 20 10 0 −2 10 10 −1 Figure 8. 10 Frequency (rad/s) 0 10 1 10 2 Problem 8. 1 zh KT z .4 gives HI z = T Kh 1 z.5 The transfer function of the integral part of the PID-controller 8. 8. The derivative part of 8. 1 = z h + T d=N . 71 . Nh + T d which is the same as the derivative part on page 308.22 has the transfer function KT s GD s = 1 + T ds=N Using backward di erence gives d KTdz . i which is the same integral part in 8.1 KT = h + T d=N Td d z . 1 d d 1+ d zhN z .23.22 is K GI s = T s i Using Euler's approximation 8.

1 +zhN=T d A small value of Td can make Hd unstable.a. The sampling interval is. Euler's approximation gives a delay of one sampling interval before an error will in uence the integral part. usually short for digital PID-algorithms. Problem 8. however. b. 1 i i This is of the same form as 8. 1 i An error will then directly in uence the computation of the integral part.6 Using the bilinear transformation gives = K 1 + 2T 1 + 2T + hz . 1 Hdz = z . Approximation of the integral part with backward di erence gives Khz HI z = T z .24 with . Euler's approximation for the derivative part gives KN . Since the D-part of a PIDcontroller sometimes is not used it is necessary that the regulator remains stable when Td = 0.

h Kd = K 1 + 2T i Tid = Ti + h=2 1 0 .

h 1 1 HT z = K B1 + 2 z . 1 A @ Ti i Ti h z + 1 . 1 C = K 1 + 2h + T z .

h .

At the desired cross over frequency we have jGi!cj = 0:525 arg Gi!c = . This gives K = 1:85 T = 149 72 .15 degrees at !c .115 We will use a PI controller of the form Gr s = K Ts + 1 Ts and we want the gain 1=0:523 and the phase . 2h Problem 8.10 has the transfer function 0:000468 Gs = s + 0:0197s + 0:0129 a.7 The tank process in Problem 2.

Tustin's approximation with warping gives with = !c = tan!c h=2 z . The characteristic equation of the closed loop system is s3 + 0:0326s2 + 0:00112s + 0:00000581 = 0 The roots are s1.0:0135 0:0281i and s3 = . The complex poles have a damping = 0:43. The zero of the closed loop system is .0:006. The sampling interval is 6 dashdotted and 12 seconds dashed.4 250 Time 500 b.1 1:85 z + 1 + 0:0067 Hr z = z.2 = . c. Figure 8.1 z+1 .0:0062.1 Output 0 0 Step response of the tank process when controlled with a continuous time solid and a discrete time PI controller.

2 for choosing the sampling period gives h 6 . 20 seconds The choice h = 12 seems to be reasonable.1 . 1 Using the rule of thumb from Section 8. 1:85 + 0:0067 1 + 0:0134 = + 0:0067z . This gives = 0:165 and ! Hr z = 1:925 1 + 0z:0778 .

73 . Fig. d. 8.4 shows simulations of the step response of the system controlled with the continuous time and the approximate discrete time PI-controller when h = 6 and 12 seconds.

e. It is possible to calculate backwards to nd out the corresponding damping and natural frequency for the controllers in a and b.1 h .h . . ` 1 .h 1 . 32 2 . : 1 .6.24 and 8.h . Using 8. BLh=2 = : 2 4 . 4h . ~ M = I . `2h . h 8 9 c. h 4 . h= 8 9 = : 2 .25 give 8 9 h= 2 ~ L = LI + A . 1 + e.5 shows the stepresponse of the system when using the continuous controller and the controllers in a and b when h = 0:25. A discrete time state space 8 9 representation of the motor is given in A. LBh=2M = 41 .e.5 Problem 8. 1 . 8.1 Output 0 0 1 2 Time 3 4 5 Stepresponses for the motor in Problem 8. `1h . 1 + e. The continuous time controller is ut = Muct . e.9 when a continuous time solid. Lxt: A discretization is obtained by sampling uc and x and letting u be constant between the sampling period points i.h 9 1 .9 a. Using L = : `1 `2 . Fig. e. gives 8 e.L = : .`21 . Lxkh b. we get ukh = Muckh .h .h 74 . a discretized dash-dotted and a modi ed discretized state dashed feedback controller is used when h = 0:25. Figure 8.

h 4 .16 we can calculate the corresponding continuous time systems. For the discretized controller L = 8 9 : 2 4 . we get 8 9 z2 . 75 8 9 . we get z 2 . .0:885 9 . We rst want to compute a state feedback such that A . we get = 0:71 !0 = 2:31 8 9 = 0:77 !0 = 1:96 The change in the damping is smaller when the modi ed controller is used and the change in the undamped natural frequency is also smaller.L = : 0:171 0:914 . 2 1 The characteristic equation of A .10 a. BL = : . Problem 8. BL is s + 3s + 2 + `2 + `1 = s2 + 5 + `2 s + 6 + 3`2 + `1 s + 8s + 32: This gives 8 9 L = : 17 3 . . BL has the characteristic equation s2 + 8s + 32 = 0: 8 9 Assume L = : `1 `2 . 1:221z + 0:442 = 0 and and for the modi ed controller L = : 2 . . 4h . then .`3 . 8 9 and for h = 0:25 and L = : 1:75 3 .8 0:336 .2 1 ` . A .L = : 0:164 0:885 . These two matrices have the characteristic equations For h = 0:25 and L = : 2 4 . we get 8 0:392 .0:664 9 . 1:305z + 0:471 = 0: From the equations given in Example 2.

Problem 8.16 gives L = LI + A . : . 5h=2 .h 1 + h .b.1 I xk = A . BL h = L : . 2h = 1:6 b. 2 8 9 ~ M = : I . = : 1 . The response when using the modi ed discrete time controller from b is shown in Fig.17 give 1 ~ L = L I + A .h=2 1h=2h . = : 0:8 1:7 .17h=2 1 . 3h=2 h=2 9 = : 17 3 . h . 3 h . Ah + KCh. : . 2 .7. Using 8. q . M = 2 .1 = This gives 1 h 1 2 : . Using the backward di erence approximation gives 1 .12 a. 1+h+h .1 xk + Bhuk + Khy k x ^ Introduce 0 = I . LBh=2 . Fig. Modifying L using 8.6 shows the output when using the discrete time controller in a for di erent values of h.8 h=2 BL 8 9 1 . 3h 3 + h . 8. h 2 . Ah + KCh^k = q . KC ^k + Buk + Ky k x h ^ I . 8. 9 8 9 8 9 8 1 h=2 9 8 = : 1 2 .16 and 8.h=2 1 . 8 9 = : 171 .

y k ^ 76 . 8 9 8 9 xk = 0xk . 1 + 0Bhuk + 0Khy k ^ ^ 8 0:81 0:16 9 8 0:03 9 8 0:19 9 = : .0:16 0:97 . uk + : 0:16 . 1 + : 0:19 . xk .

2 dash-dotted and 0. The response for the continuous-time controller is also shown solid.10 when the state feedback controller in a is used with h = 0:1 dashed . Figure 8. The response for the continuous-time controller is also shown solid.3 dotted. Figure 8.10 when the modi ed state feedback controller in b is used with h = 0:1 dashed.3 dotted.1 0.5 0 1 The response of the system in Problem 8.2 dash-dotted and 0. 0.7 2 Time 3 4 77 .6 2 Time 3 4 1 0.5 Output 0 −0.5 Output 0 −0.5 0 1 The response of the system in Problem 8. 0.

2 + sin3h!z .3 + b.1 + 2he. Flow control with local feedback. That this is the correct answer is easily seen by making long division.1 a.h Hdz = z he e. The time function t e.2 + Problem 10.h Y z = z he e.1 + sin2h! z. c.z 2 . Consider a system with Hdz = Y z The impulse response of Hd will thus be sinkh! . 2zsin !h + 1 cos !h See Table 2. Temperature control in houses.1.3 Using the model of the disturbance gives yk + m = C q wk + m: Aq Introduce the identity qm. h This can be found by looking in a table of z-transforms. The time function y t = sin!t has the z-transform z Y z = z 2 . The desired system thus has the z-transform . b. Hdz = sinh!z . Better sensors. Problem 10.1 C q = Aq F q + Gq 78 . for instance a tachometer with less noise.t has the z-transform .z 2 .h z .2 a. h Long division gives Hdz = he.Solutions to Chapter 10 Problem 10.2h z.

Then yk + m = F q wk + 1 + qGqq wk = F qwk + 1 + qGqq y k A C If wk + 1.0:5f2 + g0 with solution f1 = 0:5 f2 = 0:25 g0 = 0:125 The predictor at k + 3 given data up to and including k is thus y k + 3jk = 0:125q y k = 0:125y k ^ q Let wk be zero except at k = 0 and 5 when it is assumed to be one then k y k = 0:5y k . Introduce 8p p 9 P = : p11 p12 .1 0 0 0 1 0 1 0:5 0 2 0:25 0 3 0:125 0:125 4 0:063 0:063 5 1:031 0:031 6 0:516 0:016 7 0:258 0:008 8 0:129 0:129 9 0:064 0:064 10 0:032 0:032 where deg F = m . Since R1 is symmetric P is also symmetric. 1 + wk ykjk . Let Aq = q . 12 22 79 . 0:5q2 + f1q + f2 + g0 This gives the system of equations 0 = . 0:5.The operator qGq =C q is casual since deg C = deg G + 1. : : :. 3 ^ . wk + m are assumed to be zero then the best prediction of yk + m is y k + m = qGqq y k: ^ C Problem 10.11 we nd that the stationary variance of x ful ls stable P = P T + R1 The stationary covariance exists since the system is stable. C q = q and m = 3 then q2 q = q .0:5 + f1 0 = .4 Using 10.0:5f1 + f2 0 = . 1 and deg G = n . 1.

0:31 9 P = : . n where ai . = : .3:57 0:4 + 3:27 0:2 0:93 0:4 + 1:68 0:2 . n = 0: This is called the Yule-Walker equation.0:31 2:61 . y k . 2 0 1 2 1 . Finally 8 1:19 . : 0 0:2 .then 8 p p 9 8 0:4 0 9 8 p p 9 8 0:4 . Using the results on matrix functions in Appendix B we get = 0I + 1 where 0 and 1 are obtained from 1 = 0 + 1 1 2 = 0 + 1 2 The solution is . 1 + + an ry .0:24p11 + 0:08p12 p22 = 0:36p11 . 1 + + an yk . 0 From the state space description we get the input-output description of the process yk + a1yk .1 = . .5 . : : :. 8 9 1:19 0:4 . Thus ry + a1ry .12 1. 0:24p12 + 0:04p22 + 2 0 2. 2 1= . The eigenvalues of are 1 = 0:4 and 2 = 0:2. + : 0 12 22 p11 = 0:16p11 + 1 p12 = . n are the coe cients of the characteristic polynomial of the matrix .0:6 9 8 1 11 12 : p12 p22 .0:31 0:4 rx = : . 1 + + cnv k .1 . 9 This gives The solution is The stationary covariance function is Exk + xkT = P It remains to compute . 0:2 0:2 . and take the mathematical expectation. is independent of all the terms on the right hand if n + 1. n = c1v k .30:4 . 80 Problem 10. i = 1.0:6 0:2 . 1 2 and 8 9 0:4 0 = : . Multiply the equation above by y k . : p11 p12 .

1 + a + z + bz .1 + b = Using the spectral factorization theorem Theorem 10. Using Theorem 10.1 where z = ei! H z = C zI . 2 z.1 + b a gives the relationship Identi cation with y 2 2 21 + c2 = 1 1 + b2 + 2 1 + a2 2 2 c = 1 b + 2 a Problem 10.1.1a2 81 The process is .1 1 z + bz + b + 2 z + az + a z + az + bz .3 we nd that we can generate the same spectral density by sending white noise through H1z = z +za+zc + b this gives the spectral density 1 2 . 9 1 z+b 9 .1 1 = 8 1 : z+a z + b. s.2 we nd that the spectral density of y is y = H z v H T z . From Example 10.11 9 z +a 2.1 + b z z.1 + az .: 0 2 0 .Problem 10. 8 2 09 1 v=: 0 2.: 0 2 1 0 2 9 8 .3 with the exception that Ev kes = r12 k . 8 1 = : z+a y 1 z +b 2 98 .7 xk + 1 = axk + v k yk = xk + ek This is the same as in Example 10.1 2 .1++cz + bcz .3 we know that r rx = aj j 1 . = : 1 and Thus 8 98z +a 1.1 + = 2 z + az . The covariance function for x will thus be the same but the covariance of y will contain an additional term due to the correlation between v and e.1 +b 2 1 = z + az .: 1 . .6 There are two noise sources v1 and v2 that is used to generate y k.

a 2 i! . + re where it has been used that rex = rxe .a r1 1. r12aei! .1r12 + 0 + 0 1.i! . 1 ea i! . ra + 1 . .i! = y ! = 2 y =.i! ei! . a . where it is r12. 1 . a a 1 where it has been used that 1 1 X r e. It is natural that white noise in the future is uncorrelated with the present value of x.a2 8 a. a . ce. The rst term is zero for 0. c 2 ei! . 2 r r 12 12 = 21 1 . 2 aj je.1 The spectral density for is see Theorem 10.The covariance of y is ry = E fy k + ykg = E f xk + + ek + xk + ek g = = rx + rxe + rex + re = rx + rxe + rxe. a 2 82 . a e.1a2 .1a2 + 0 + 0 + r2 : a r1 2 + a .i! .1r12 and 80 0 = 1 = 0 .1 r12 + 0 r ry = 1.a+ r2ae .2 1 X q. a "k y ! = 2 ei! .i! = ei! . a = 21 r1a + r121 .1a2 + aj j.1a2 .1 . in stationarity. ra ei! .1 r a 12 1 0 1 0 =0 = 0 + 0 + a.1a2 + r2 = . r aj j 1. a a . 1 ea i! . =.i! .i! .e. This gives rxe = : r12 a .c y k = q .a2 + r2 The de nition of spectral density gives = 6= 0 =0 r . rxe + 1 = E fxk + + 1ekg = E f axk + + v k + ekg = = arxe + rve The last term is zero except for = 0. ae .1r12 r1 1. ae.

1 + r H z + r H z . 2ar = 21 + c2 2 1 12 8 98 9 : H q 1 . 1 where H z = z.1 = 21 : H z 1 .1 + r = 2 1 12 12 2 .a . y 1 12 2 1 r H z H z .2c r1 + r12 1 . The spectral density of y is given by 8 9 r . r12 1 + a + r2a 1 + a = 21 + c2 a a a 2 2 2 r2a . y k = ek 8 98 . : H z . : vk . r2a = . : r 1 rr12 . r12 = 2c A more elegant solution The ouput can be written as r 1 + a2 + r .Identi cation of 1 with 2 gives the relation 8 :r12 . a .

r2a + r1 .11. a + r2 . r12 + aj j. a ek = q . 2zar12 zr.c a. a + r12 z .7 we nd that r1 = a . 2ar12 = 2 1 + c2 ar2 . a + z . r2a .c yk = q .8 The process can be written as q.1 + 21 + 2 = 21 z r12 .1 . aa + z r12 . c r2 = 1 Thus aj j ry = 1 . 1 r1 = 21 z . c2 r12 = a .1 .a r21 + a2 + r1 .1r12 a 83 . a ek + ek = xk + ek where xk + 1 = axk + a . az . cek Using Problem 10. r12 = 2 c 9 which gives the same equations as in the previous method Problem 10.

The formula for I2 gives B0 = 1 + 0:22 + 0 = 1:04 B1 = 20:2 + 0 = 0:4 B2 = 0 e1 = 1:7 and 17 1:5 ry 0 = I2 = 1 .c 2 ry 0 = 1 .1:5 1 2 = 0:7 0:51 .1a2 + r2 = 1 .. c = 21 1:1:25 .4.1 = 21 z . 84 .0:8824 .22ac = 1:08 a Further 10. 1 4 cos where z = ei! .0:45 0:1129 1:1765 0:1129 This gives I2 = 10:4167 1:1765 + 0:3922 0:2 12:33 These calculations have been performed in high precision and thereafter rounded.1 .4 can also be used to get 1 . 1+50:+q0:7 ek = a 0q2 + a1q + b2 ek :q q+a 0 1 2 can be determined using Theorem 10. cz .1:5 1 0:7 .0:45 0:51 0:51 1 = .:cos ! ! y . 0:1:04 :7 :.4. a 49 .9 The variance of the process b q2 + b q2 2 yk = q2 .17 gives z .0:45 1 0:2 .with a = 0:7 and c = 0:5 we get for 6= 0 02 0 ry = 0:7j j 1 . az . Problem 10. 2 ry 0 = I1 = 1 +1c. +:50:4 :51 . a + 1 = 1:08 2 The variance can also be obtained from Theorem 10.1:5 0:7 1 0:2 0 0:7 .:0449 + 0::7 0:7j j = 0:360:7j j 0: For = 0 r a. 0:7 = 12:33 491 1 1 2=0 1 = 0:3922 0 = 10:4167 The recursive scheme in Section 10.

b2 cos ! 2 Identifying with the desired spectral density gives 2 1 + a2 = b2 1:36 2 2a = . a = . 12 + 9ek .4 = .11 a. i! 2 . 2ek .a e 1 + a2 . 2 . 1 y = 1:36 + 1:2 cos ! Using Theorem 10.4 = .i! 1 2 b2=2 . 4ek . 32 = E ek2 + 4ek .2 we get b H z = z .i! e . 4ek . 2 . 22 + 16e6 . 1 + 3ek .2 .b . 3 ry = Ey k + y k This gives ry 0 = E ek .1:2b2 85 .20 ry 2 = 1 3 + . 2 .2 3 + 3 .i! = i! . 32 + crossterms = 1 + 4 + 9 + 16 = 30 The mean value of the crossterms are zero since E fek + ekg = 0 In the same way 6= 0 ry 1 = 1 .2 + . 1 + 3ek .4 ry 4 = 0 k 4 Problem 10.Problem 10.i! = H ei! 2 H e. a 2 e + e = b2=2 1 + a2 . 2a cos ! = 2 2 1 +2a b 1 2a . 2ek .4 = 11 ry 3 = 1 .10 The process is y k = ek . a 1 y ! = H ei! u ! H e.

1:36 a2 + 2:72 a + 1 = 0 2a 1:2 1:2 r 1:36 1:36 2 . Problem 10. 86 .4 CCS Var y = I1 = 1 . 1 = . 5:40 cos ! = 2 1:81 . + : 1 0 .1:2 2 = 2 The desired lter is 2 b. = : 0:3 0:2 . :62 = 0264 0 : H z = z +20:6 p Problem 10. 2a cos ! In this case r1 = 1. Identify with the desired spectral density gives 1 3 1 1 2 5:43 .0:6 a = .12 The stationary covariance is given by 8 0:3 0:2 9 809 xk + 1 = : 0 0:5 . 1:2 1:2 r 2a p b = . : p11 p12 . p22 0 0:5 p12 p22 0:2 0:5 0 0:5 p11 = 0:09p11 + 0:12p12 + 0:04p22 + 1 p12 = 0:15p12 + 0:1p22 p22 = 0:25p22 + 0:5 8 1:1385 0:0784 9 P = : 0:0784 0:667 .4 shows that the lter b H z = z . 1:8 cos ! This gives a = 0:9 and b = 1. xk + : 1 . uk + v k 8p : p11 12 P = P T + R1 9 8 98 98 9 8 9 p12 .Divide the two equations 1 + a2 = . : 0:3 0 . Theorem 10. a gives the spectral density r1 b2 ! = 2 1 + a2 .13 Example 10.

at.at.8. kh = Zt kh b e.kh The discrete time loss function is obtained from CCS 11.1 For the system we have d t.t. kh = 1 t. which gives Q1 = Q12 = Q2 = Z kh+h Zkh +h kh .Solutions to Chapter 11 Problem 11.at. e.s b ds = a 1 . kh = . kh.kh and . dt This di erential equation has the solution kh.6-11. kh = e.at.

kh2 b = a2 Zkh +h .

Problem 11. 2a3 3 . :1. 4e.2ah Notice that there will be a Q12 term even if Q12c = 0. e.= 1 x+ 0 u 0.ah 2 1 .2 80 x = :0 _ 8 y = :1 = eAh . e. 87 h.kh 2 + e.as. 9 0.2ah ds a2 b2 + h .x 9 8 9 Sample the system Zh 0 81 = :0 eA 8 h2=2 9 Bd =: h .as.2as.as.kh ds = 21a 1 . 1 9 .ah + e. b2 kh b b e. e.kh a 1 . e.kh ds = 2a2 1 .

1..T S . The loss function is y 2 + u. h 0 0 Zh 8 1 Q1 = : 98 9 81 : Problem 11. ::. 0 0. ab xk = . 1 0 0 ! h 8 8 2=2 9 ! Zh Z 9 3 5 4 2+ 2 =2 .e.1.Lkxk = .17 gives + 12 sk = a2sk + 1 + 1 .Q2 + . Problem 11. a2b2 bs2skk + 1 = 1 k = N .Sample the loss function Q1c = I Q2c = 1 Q12c = 0. d = : h4=8 + h2=2 . d = : 3=2 + .6-11. 0 0 h8 9 8 2=2 9 Zh 8 2=2 9 8 h3=6 9 Z Q12 = : 1 0 . : .8 we get 9 8 h 9 h2=2 2 + 1 . T S . : . d = : h2 =2 h3 =3 + h .T S Let 88 8s s 9 S = : s11 s12 . i.:0 1.3 The Riccati equation11. 1 and 11.2 ba = a b The minimum loss is given by and the closed loop system is uk = .5 a. +1 d = 1+ d = h+ h + h Q2 = : 4 3 20 Z 0 1 d = 1.19 gives which gives the controller Lk = b. 12 22 . The steady state value of the Riccati equation is obtained from Q1 = C T C 81 = :0 9 S = T S + Q1 . a xk b2 b min J = x02s0 = x02 xk + 1 = 0 The state is zero after one step and the resulting controller is thus a dead-beat controller. Using 11.

1 + K = 0 There is one pole in the origin and one in 1 . K . K = 0 and as ! 1 then 1 .5 0 −2 10 Figure 11. s12 = 1 + p1p 4 + K = + s 22 2 + 1 + 1 + 4 b. +12 s 22 22 22 s12 = s11 .K .1 shows how the pole varies as a function of . Fig 11. 89 . For = 0 then 1 .5 as a function of the control weighting .L = : . +12 s s2 12 s2 s22 = s11 . 1 + K + K = .K . K ! 1.1 Pole 0. The poles are obtained from . The dynamics of the closed loop system is 8 1 9 1 . . 10 10 10 Control weighting rho 0 1 2 10 3 10 4 For the inventory model we get s2 s11 = s11 + 1 .1 10 −1 The pole in Problem 11. + s The solution is The feedback vector is where 1 s12 = s22 = 1 + 2 + 4 s11 = 1 + s12 p 8 9 L = K : 1 1 .

u(k) 2 0 −2 0 5 10 c) y(k).T S .1 z 2 . = rp1 2 + 1 = r1 + p2 1 which has the solution r = 2 + 1 . Figure 11.40. For the inventory system Az = qq . 1 + 4 2 It is easily seen that this is the pole in 1 . b = 0:5. z + 1 = rz2 + p1 z + p2 z . 1 + 4 2 The poles of the closed loop system are thus one at the origin and one at 2 2p p .2 .a) y(k).5 when a = 0. z . = rp1 + p1p2 2 + 1 = r1 + p2 + p2 1 2 Since r = . u(k) 2 0 −2 0 5 10 b) y(k). 1 B z = 1 and we get z. p 90 . K .1 + p2 or 0 = rp2 .2 + p1 z. + Q2 6= 0 then the rst equation implies that p2 = 0 and we get . 1 + 4 = 2 + 1 +2 1 + 4 p p1 = .p1 = 2 + 1 . c = 5 and d = 25. u(k) 2 0 −2 0 5 Time 10 d) y(k). u(k) 2 0 −2 0 5 Time 10 The output solid and the control signal dashed for the system in Problem 11. 2 + 1 .2 The poles of the closed loop system can also be determined from 11.

1 :026 0 030 + z2 :. 2:7720z 3 + 4:1849z 2 . sampling of the loss function and solving the Riccati equation.1 = 0 This gives . Problem 11.02:030z65z+10+ 0:68 65z + 0:68 . 1:z + 0:026 10. 1:z + 0:026 z .e. 11. 1: .6 when is varying. 1:65+ +:026zz 2 = 0 65z + 0:68 z 0:68 0:00078z 3 + 0:001576z 2 + 0:00078z + 0:68z 4 . 1:z + 0:026 65z + 0:68 Only the output and the control signals are penalized in the loss function. The following Matlab-macro illustrates the solution of Problem 11. The closed loop system has poles that are determined by the stable roots of Problem 11. like Control toolbox for Matlab. 0 Re 1 The system has the transfer function 0 030 H z = z2 :.9 Solving LQ-problem for system of higher order than one by hand causes much work.1 Im 0 −1 −1 Figure 11.9.3 The closed loop poles in Problem 11. the design is signi cantly simpli ed for the control engineer. sampling of the system. With numerical packages. 2:7720z + 0:68 = 0 Fig. or 2 :030z 0 0 030 = + z 2 :. 91 .3 shows the closed loop poles when is varying. i.6 + H zH z .

Transform continuous-time LQG problem to the corresponding discrete-time LQG problem Phi. With A z = z 2 .08.Q12. z + z 2 + p1z .Gam. h=5.Q1.5 we have determined r. k .Q12c. Q2c=rho.B.0005. Q12c= 0.Q2. 0 0 .e. Linear quadratic regulator design for discrete-time system L. 0 .Q12.37 may be used to get the exact values of the gain margin. z P z = z2 + p1z we get z 2 .Q2.10 In Problem 11.h. 11.2. L The design gives L = 3:055 108:7 Problem 11. 0 -alpha .1 max and max are also shown in Fig.Gam.R1. 1 = 0 I.001.Q2c. A= 0 1.Q1. Q1c= 1 0. 1 1 4 min = 0 . B= 0..zeros2.S = lqrdPhi.9 CCS alpha=0. the system is stable if p1 + .4.1 + p1 + 4 = min .Lv. rho=0. then 2 = r 2 + 1 + p1 + 4 Equation 11.Macro for Problem 11. k=0.Je = lqgsampA.Q1c. z 2 + z = z z + p1 + . 92 .

r1 = r2 and r1 r2. xkjk .37. 93 . Finally if r1 r2 then P 1:33r1 and .e. 1 = 0 ^ x ^ ^ 5 K k = r0:+PPkk 2 2 : : P k + 1 = 0:25P k + r1 .5 shows .11 has a control signal uk. Problem 11. In the rst case P r1 and . KC 0.11 xk + 1 = 0:5xk + v k y k = xk + ek Theorem 11. P 0 = r0 r + The dynamics of the lter is determined by r . KC 0:5. the system is xk + 1 = 0:5xk + v k + uk yk = xk + ek Determine a steady-state LQG-controller when Q1 = 1. KC for di erent values of r1 when r2 = 1. r1P = r1 r2 Consider three cases r1 r2 . i. x0j .10 Gain margin 5 0 0 Figure 11. Q12 = 0 and Q2 = .4 5 The gain margin 10 Control weighting rho min dashed and max solid from 11. The system is Additional problem Suppose that the system in Problem 11. In the second case P = 1:13r1 and . 0225PPkk . KC = 0:23.5 gives that the Kalman lter is de ned by 8xk + 1jk = 0:5^kjk . K k = r 0:5P2k 2+ The steady state variance is given from P 2 + 0:75r2 . 1 + K kyk . Fig. 1. 11. KC = 0:5 .

17 and 11.12 a.L . .0. 1 + . C xkjk .11 for di erent values of r1 when r2 = 1. 1 + K y k .47 gives P k + 1 = P kT + R1 .0:75 + 1 +p 0:75 . Figure 11. Equation 11.LK + L 5+K Problem 11. C xkjk . 1 + 4 2:5 + 2 + 2 0:75 .1K Y q Hreg q = q . P kC T R2 + CP kC T . 0:.L + KC .5 Pole 0 0 The pole of the Kalman lter in Problem 11. 1 + Ky k x and thus or U q = . 1 ^ x ^ = ^kjk . 12 + 4 p Using the Kalman lter from Problem 11. + .19 gives S = 0:25S + 1 .11 and the LQ-controller gives xk + 1jk = ^kjk .Lxkjk . 1 + .1CP kT 94 . 1 x ^ ^ = .uk + K y k . .5 5 State weighting r1 10 Solution to the additional problem Equation 11. KC ^kjk .LqI . 0:25S +S 0:5S L= +S which has the solution 2 2 L = .

Fig. KC = 2 = 0 The lter has a dead beat response. p12kp11k + p12k = p11k + 1 p11k 2 p22k + 1 = p22k + 1 . 1 = : 1 1 . 95 . The steady state value of P is 81 19 P = :1 2. p11kp +p12k 11 k 2 = p11kp22kk. p k = : 1 .6 The elements of the variance matrix in Problem 11. p p12k = 1 + p11k + 1 11 Further 8 k 9 8 p k + p k 9 1 8 2 9 K k = : k1 . . R1 = . b. x0j .4 p11(k) 2 0 p12(k) 0 Time p22(k) 5 4 2 0 4 2 0 0 5 0 Time 5 Figure 11. with 1 1. 8 9 T = :0 0.v .12. = : 11 p k12 . The initial values of the lter are For k = 0 K 0 = 1 0 T i. K is timevarying. 11 2 12 k 0 The poles of the lter are found from detI .e.5 shows the elements of the covariance matrix as a function of time.v 8 90 1 C = :1 0. ^ 8 9T and assume that P 0 = 3I . 11. 2 81 = :0 9 we get p11k + 1 = p11k + 2p12k + p22k . p12k p11 p12k + 1 = p12k + p22k .

q 1 v k + ek . and .1 vk + 2 + . Using Theorem 10.2 = : 1 .50. 1 ~ + 96 .2 + . 1 v k = q.a q.1 . q. 1.1 .T 0 1 1 81 39 9 1 1 0 + 0:01 5 3 = : 0 1 .a q .Problem 11. v k +1 and vk +2 are independent of y k then the best estimate of xk +3 is given by xk + 3jk = 3 xkjk + 2 + + I . q yk . The state at time k + 3 can now be determined xk + 3 = xk + 2 + .1 v k + 2 + . ! 1 .xk = 1q. y k ^ ^ xkjk.T 28T + . P kjk : 3 1 . .15 xk + 1 = axk + vk y k = xk + ek We use the exponential smoothing estimator cov v = 1 cov e = xkjk = xk .2 + . xk ^ .1 v k + . P k + 3jk = 3 P kjk3T +8 :019 1 . . aq 1 v k + 1q.1 v k + 1 + . Since v k. 1 + 1 .1 v k + 1 + .2 + . :3 3. 8 9 Problem 11. q ek . ^ ^ 81 2 = :0 3 xkjk + 4:5 :3. .4 we get 2 2 var xkjk = 1 + a1 + 1 .T T + . = . a + 1 . . 1jk . ^ 9 8 9 The variance of the estimation error is 2.05. x0 + 4:5 ^ and the variance of the prediction error is 0.2 The best estimate of xk given y k is determined from 11.14 Introduce the notation 81 19 809 8 0:5 9 = : 0 1 .1 = : 1 . q .2 = 3 xk + 2.1 v k + 2 + 2.2 = 2 xk + 1 + . If x0 is known then P 0j0 = 0 and y 3 = : 1 3 .

uk + : 0 . 1 + a2 + 1 min = a21 + a + a . 1 + K yk . smoothing: Kalman: = 0:4222. X k + : 0 . K = 0:2650. a1 + a + 1 . KC ^k .P + = . v1 k 0 0 8 0 0 19 : yk = : 1 0 1 . Ev1 = 0 ek = e1k + me . a = 0:5. a22 2 2 The gain in the Kalman lter is 2 var xkjk = P . 1jk . a2P . use Maple. 1jk . 1 Kalman with direct term p xkjk = ^k .1 . 1 + Ky k x This will have the same form as the exponential smoothing estimator only when a = 1.16 The scalar state equations are xk + 1 = axk + uk + v k y k = xk + ek vk = v1k + mv . Kalman variance a2 P 2 P = a2 P + 1 . a + 1 1 + 2 1 + a2 + 21 . 1. C ^k . P + This gives . 1 . X k + e1 97 8 : x mv me 9T . Ee1 = 0 Let and 8 X k + 1 = : 0 1 0 . 1jk . 1 ^ x x = I . P P = PP+ = a K ~ + K = PaP + Numerical values: = 1. X= 819 8a 1 09 819 .Minimize with respect to . Exp.a2P 2 2 p P = 1 . var xkjk = 0:6181 ~ var xkjk = 0:5311 ~ Problem 11.

= : a 1 a2 a + 1 CA2 19 1. It is.lter can be designed.17 The constants a1 and a2 are determined by the following two conditions 1. xk2 = E xk . The correct mean value should be obtained. The observability matrix is with 8 C 9 81 Wo = : C . 1me + mv gives 8a 19 819 819 8 X k + 1 = : 0 1 . however. a1e1k .The observability matrix is then 8 C 9 81 0 Wo = : CA . Problem 11. 18 = 0 a1 = 10 The estimator is thus 1 9 xk = 10 y1k + 10 y2 k ^ 98 . = : a 0 1. a 2 9 = 10a2 + 9 . 9 from which follows that rank Wo = 2. a1 xk . X k + : 0 . a2xk . The variance should be minimized. Rede ning the state vector X as 8x 9 1 X = :m. 1 rank Wo = 2 This means that me and mv are not both observable and no Kalman. a2 e2 k2 ^ 2 1 + a2 9 = a2 + 1 . Condition 1 gives that a1 + a2 = 1 The variance of the estimation error is V = E xk . where x1 = x + me m = a . possible to design a second order observer with reconstruction of a linear combination of me and mv . v1k 8 9 : yk = : 1 . uk + : 0 . X k + e1k 0 Reconstruction of x1 and m is possible if these states are observable. 18a = a1 1 1 2 1 1 Taking the derivative with respect to a1 gives the condition 9 20a1 . 2.

0:45 9 819 xk + 1 = : 1 0 . This gives 9 xkjk = 10p p 9 y1 k + 10pp+ 9 y2 k + 10p9+ 9 xkjk .1 = 0 P z = b1 zB z = 2z 0:5z + 0:38 = z z + 0:76 1 Now we have to nd L such that 8 . . In the example R1 = 0 the steady state gain will then be zero since the estimate will approach the true value of x. 99 8 1:45 .1 = rP z P z . An alternative solution is to use 11.1 = 10pp+ 9 : 9 1 . where controllable canonical form have been used. i.54 we get P 1j0 = p and 8 9 .L = : 1 0 . Assume that the a priori estimate of x is zero and the variance of x is p.0:45 .9 Vmin = 10 Using only y1 gives the variance 1 and only y2 the variance 9. Using Matlab we get the solution 8 0:25 0:19 9 S = : 0:19 0:1444 . The minimum value of V is Problem 11.40 AzAz. a combination of the measurements gives a better result than using only the best measurement. This gives 9 8 L = : 2:21 .17.50 11. Thus. x0j0 and P 0j0 = p ^ From 11.1 + Bz B z.0:45 . K 1 = P 1j0C T R2 + CP 1j0C T . xk + : 0 . uk 8 9 yk = : 0:5 0:38 .0:76 0 9 . Q1 = C 0:19 0:1444 . xk 9 8 T C = : 0:25 0:19 .20 Q12 = 0 Q2 = 0 The steady state solution is obtained from 11.e. 8 9 L = : 2:21 . 1 ^ ^ + If p is large then the weights for y1 and y2 will be those calculated above.

Lk S k + 1 05 Lk = + :SSkk 1 + This gives .1 gives the Riccati equation S k = 0:5 0:5 .Problem 11. S N = S 2 = Q0 = 1 0 52 S 1 = :+ 1 53 L0 = + 0:+ 0:52 1 : L1 = 0+51 The control law is uk = . First assume that = 0 and use linear quadratic design with Q1 = Q12 = 0 Q2 = Q0 = 1 N =2 Theorem 11. In this case = 1 and xk is reconstructed using a Kalman lter xk + 1 = 0:5^k + uk + K k y k . Theorem 11.21 a. 0:5P kK k with P 0 = R0 = 1. For di erent values of we get 1:0 0:1 0 L0 0:056 0:093 0:1 L1 0:250 0:455 0:5 b.Ly k.5 gives R1 = R12 = 0 R2 = = 1 . xk ^ x ^ with .Lkxk = . This gives k P k K k 0 1 0:25 1 0:125 0:056 100 . R0 = E x2 0 = 1 5 K h = 10:+P k Pk P k + 1 = 0:25P k .

xk + ek 8 2 09 R1 = 0:01. R2 = : 01 2 .The control law is uk = . xk + 1 = xk + v k 819 yk = : 1 .22 a. 2 8 1 9 .Lu^k. x Problem 11.

xkjk . 1 + Ky k ^ ^ 8 9 . xk + 1jk = 1 . K : 1 .

1 2 :1 1. = 2 2 + p 2 + 2 p 2 p 1 2 2+p 8 9 0:09458 8 91 2 = : 4 1 . 8 2 29 : 2 1. :1. 11.8 12 0 9 8 1 9 8 9 . 8 98 2 p 9. p2 2 2 1+ 2 2 2 2 2 = 0:01 1 2 + 1 + 2 p 2 2 p2 . :1. : p:1 1. 0:01 2 1 2 2 = 0 1+ 2 c.1 8 1 9 = 0:01 2:1 1. :1.: K=p .: 1 + p p .1 8 1 9 :1 1. 0:01p . p = 0:005 0:0052 + 0:01 2 1 2 2 1+ 2 r = 0:005 + 0:0052 + 0:01 4 = 0:09458 5 s 2 2 101 . 8 9 8 1 + p p 9.47 p = p + 0:01 . 2 + 0 2 . 4 + 5 0:09458 = : 0:0846 0:0211 . 2 p 2 +p p2 b.

0: 21 + 0 4 + 1:2 4 1 2 = 4:28 This implies that the prediction variance is almost the same as the variance of y when m 3.1 The m-step ahead predictor is given by where G is obtained from the identity 12. q and the variance of the prediction error E y2k + 1jk = 2 = 4 ~ For m = 2 qq 2 .:1:2 1:2+2 00::51:2 :. For m = 1 we get q 2 . 0:2q . 1:2q + 0:4q2 + f1 q + f2 + g0q + g1 which gives F q = q2 .1+4.0:14q + 0:08 and E y 2k + 2jk = 21 + f12 = 4:16 ~ For m = 3 we get q 2q 2 .Solutions to Chapter 12 Problem 12. 1:4q + 0:5 = q2 . 1:4q + 0:5 = q2 .0:41+0:4 = 1 .19. 1:4q + 0:5 = q2 . 102 yk + mjk = qGqq y k ^ C . 0:2 Gq = .17 and the variance of the prediction error is given by 12. 0:14 Gq = . 1:2q + 0:4q + f1 + g0 q + g1 This gives F q = q .0:088q + 0:056 2k + 3jk = 2 1 + f 2 + f 2 = 4:24 Ey ~ 1 2 Using Theorem 10.4 we can compute the variance of y to vary = 2 2 2 : : 0 5 22 1+1:4 +0:5 1+044+2. 1:2q + 0:4 + g0q + g1 which gives Gq = g0q + g1 = .1::4.0:2q + 0:1 The predictor is then given by q2 0 1 y k + 1jk = q.0:21:4++ :0:q5 y k ^ 2.

am. a f2 = .1 + g0 f1 = c . . a = 1 .am. The C -polynomial has a zero outside the unit circle. 1 = 5ek + 0:2ek . .2 c . aq yk ^ c and the variance of the prediction error is E yk + mjk = 21 + c . 0:9yk . Example 12.a2c .2 + ::: + fm. a2 + + a2m. 0 = afm.17 is qm. a f3 = . a . a2 1 . a2 ~ ! 2m.ac . a g0 = . It is replaced by C z = 5z + 1 = 5z + 0:2 The equivalent process is thus y k . 1 b. .1c . 0:9q + f1 + g0 This gives F q = q + 1:1 Gg = 0:99 103 . a2 + a2 c .1 shows how the C -polynomial should be changed. a The m-step ahead predictor is m.1 + f1qm.3 a.Problem 12. fm.1 + g0 This gives The solution is c = a + f1 0 = af1 + f2 . a2 Problem 12.1 2 1 + c .2 The identity 12. The two-step-ahead predictor is obtained from q q + 0:2 = q .2 + fm.2c .1 0 = afm.1 y k + mjk = . .1 = .1 q + c = q + aqm.a q +c .

m 1 2 3 4 104 y7 + mj7 ^ 4:5 3:7 3:0 2:5 zd 7 + m 380 390 400 410 z 7 + mj7 ^ 384:5 393:7 403:0 402:6 5 6:1 6:8 7:2 . 3 and 4 step ahead predictors of y.5 330 4 330 . .10 340 5 350 0 350 6 370 10 360 7 375 5 370 The prediction of the demand for August through November is z 8j7 = zd8 + y 8j7 ^ ^ . zd is the deterministic part of z . y k. 2.17 and give m 1 2 3 4 The prediction is F q 1 q + 0:7 q 2 + 0:7q + 0:59 q 3 + 0:7q2 + 0:59q + 0:48 Gg 0:7q + 0:1 0:59q + 0:07 0:48q + 0:06 0:40q + 0:05 y k + mjk = qGqq yk = g0y k + g1y k . 1 ^ C which gives the predicted values and their standard deviation .4 Using the data given up to time k = 7 it is possible to calculate y k and zd k = z k . k zk y k zd k 1 320 10 310 2 320 0 320 3 325 . .This predictor is and :99 yk + 2jk = q0+ 0q2 y k ^ : E y 2k + 2jk = 251 + 1:12 = 55:25 ~ Problem 12. z11j7 = zd11 + y11j7 ^ ^ We thus need the 1. Those are given by solving the identity 12.

0:25 + f1 = 1:55 g1 = . 0:7 k . by the stability triangle. f1 + g0 0 = 0:5f1 + g1 0 = g2 Solving these equations f1 = 1:8 g0 = . BGFq yk = q +:555q + :1:q y k q 0: 8 The loss function is Ey2 = 0:521 + 1:82 = 1:06 The noise sequence has a non zero mean value. deg B = 2 The Diophantine equation is q d. 1 = u + uk .0:9 g2 = 0 The minimum variance regulator is obtained from 12. 1 + u + 0:6 ~ Choose u = .5 The polynomials are A = q3 .Problem 12. 1 ~ = uk .6 .27 . This is a necessary condition for the minimum variance design method. and uk = q 0:10q:2 y k . 0:5y k .1 q 2 + 0 9 q uk = . Introduce ek = 2 + k uk = u + uk ~ where k is zero mean white noise. 2 + k .1 + f1 0:25 = 0:5 . 105 Problem 12. 0:72 + k . q2 + 0:5q B = q + 0:5 C = q3 + 0:8q 2 + 0:25q It is easily seen that C is a stable polynomial. e.1 C q = Aq F q + Gq qq3 + 0:8q2 + 0:25q = q 3 .g. The identity gives q u = q 0:10:2 y k ~ . 2 + 2 + k . 0:6 .0:6 and the problem is reduced to the standard problem. The pole excess is d = deg A . The process is now yk . q 2 + 0:5qq + f1 + g0q2 + g1q + g2 Identifying coe cients of powers of q gives 0:8 = .

a yk c Problem 12.2 and Fig. The output is drifting since the A-polynomial contains an integrator.Problem 12. 12.. 12.3 shows the output and the control signal for the same noise realization when the minimum variance controller is used with d = 1 and d = 2.2 1 2 ! = 2 1 + a2 + 2a cos ! This implies that = 1 and a = 0:6 gives the desired spectral density. a Gq = aa . This implies that uk . Fig. The process is then a moving average process of rst order. The process is now described by . The expression above gives the optimal controller uk = 0 if a = 0. qa+a. a. 12.8 b. 2 cannot be used to decrease the variance of y k. The identity gives F q = q + c .7 a. cq and the minimum variance controller is cq uk = .1 shows the output of the open loop system for one realization of the noise sequence.9 and the minimum variance in the two cases are d=1: Ey 2 = 1 d=2: Ey 2 = 1 + 3:22 = 11:24 b. 1 H z = z + a Sending white noise ek through this lter gives the spectral density see Theorem 10. Fig. The identity gives for d = 1 and for d = 2 F q = 1 Gq = 3:2q + 0:2 Problem 12.

y k = 1 . BGFq y k q a. 0:3 y k = q + 0:6 uk + q ek F q = q + 3:2 Gq = 5:64q2 .1 q +10:6 ek + 1 uk q or q 2 + 0:1q . Assume that 106 . 2:24q The minimum variance controller is q uk = . 01:5q.

1 50 The output of the open loop system in Problem 12.0:5 K 1:5 107 .200 Output 100 0 0 Figure 12.q + 0:6 K y k + q ek yk = q2 + 0:1 + K qq + 0:6K .8. 0:3 y k = . 100 Time 150 200 b.Ky k q 2 + 0:1q . 0:3 ek The system is stable if . Use the controller This gives uk = .2 50 The output and the control signal when d = 1 and the minimum variance controller is used for the process in Problem 12.8. 100 Time 150 200 10 Output 0 −10 0 50 100 150 200 20 Input 0 −20 0 Figure 12.

2:009. 0:6K 0077+ 0066K 2 .Kyk we get the closed loop system q 2 + :5q yk = q2 + K .10 Output 0 −10 0 50 100 150 200 20 Input 0 −20 0 50 Figure 12. 0:1 + K 2 : : K dI2 dK = 0 This gives the third order equation 72K 3 + 12K 2 . The minimum value of I2 is obtained from Theorem 10. With the proportional controller uk = . 1.004.4 gives an expression for the variance of a second order process. 100 Time 150 200 For K = 1 we get I2 = 3:87. The minimum variance is Ey 2 = 1 since d = 1. 266K + 1 = 0 which has the roots K . The value of the variance is I20:004 = 1:12 d.10 a. Only the root K = 0:004 gives a stable closed loop system.3 Same as Fig. 0025q + 0:5 ek : 108 . c.839 and 0. 12.2 but when d = 2. Problem 12. : + : I2 K = 1:3 .

4:25K + 3:25 = 0 with the solutions K = 1 and K = 3:25. deg B = 2 deg Ao 2 deg A . deg Bm deg A . This means that the process zero can be cancelled without problem.0 2:125 . 1:5q 2 + 0:7q Bq = q .2 we nd that the closed loop system is stable if 0:5 . 2 . 2 = uk . 0:5uk . K = 0:51:75 . b. 1 109 . With minimum variance control we would get E y 2 = 1.11 y k . 3 + v k gives Aq = q 3 . 1 + 0:7y k . Problem 12.1 + K . K + 0:25 or .KK . From Example 3.5 is stable and well damped. 0:5 .1:25 K 1:75 Both K = 3:25 the second root in a. 0:25251:5:.4 gives B0 = 1 + 0:52 = 1:25 B1 = 21 0:5 + 0:5 0 = 1 B2 = 0 e1 = 1:5 1: 1 5 0:25 I2 K = 1 .Using the results in Theorem 10. and K = 2:125 give an unstable closed loop system and the calculation of I2 is no longer valid. deg B + . deg Am . 1:5yk . The degree conditions gives deg Am . 0:25 0:5 . K 1:25 + K Taking the derivative of I2 and putting the derivative equal to zero leads to the equation K 2 .1 .:2521 . . 0:5 Note that the process zero 0. K = 1 gives the variance and I2 1 = 4 3 This is minimal variance for the present control law.

1:5q2 + 0:7q q + f1 + g0 q2 + g1q + g2 110 . 1 the equation becomes z 3 . 1. q1+ 1q5q1. 0:2q2 The minimum variance control law can be written as q uk = . zS z = AmzAoz Recalling the condition deg S = deg A .a.0:7r1 = . which in this case is deg F = d . v k = ek . Rq yk = . v k = 0. BGFq y k q where the polynomials F and G are obtained from q d.05:q5 q : 0 b. 0:2q2 = q3 . Minimum variance The polynomial C is given by C q = q3 .1 C q = Aq F q + Gq. Rq y k q q where Rq = R1qB + q = q + 1:5q . : uk = . q = 1 0 Bm q = 1 Ao q = q 2 The polynomials R1 and S are obtained from the Diophantine equation Az R1z + B . deg G = 2 qq3 .1:05 s2 = 0 S T uk = Rq uck . 0:5 B . 0:5 T q = Bm Ao = q 2 Assuming that uc k = 0 gives S :55 2 . 0:7 = 1:55 s1 = . Deadbeat Control Amq = q 2 B + q = q . 1 = 1. 0:2ek . 1:5z 2 + 0:7z z + r1 + s0 z 2 + s1 z + s2 = z 4 with solution This gives the regulator r1 = 1:5 s0 = 1:5r1 .

1 Rq . 0:2q . 0:3q . The variance of the output is then simply calculated as E y 2 = 1 + 1:32 + 0:32 2 = 2:78 In the minimum variance case the output is 2 y k = q. 12.1ek which also is an MA process.1 111 . Fig.d.1 1 + 1:5q .0:91 g2 = 0 The minimum variance controller is thus given by :25 2 . The system can then be written as C yk = B1 uk + A1 e1k A1 1 12. Problem 12.1ek = 1 + 1:3q. 1:5f1 + g0 = 0 0:7f1 + g1 = 0 g2 = 0 with solution f1 = 1:3 g0 = 1:25 g1 = . The output is in the deadbeat case given by CR1 yk = ARCRBS ek = A A ek + m o = q .2ek which is a moving average MA process.1 ek 1 = 1 . q1. 0q5q0+91:q3 yk : 1 c.0:2 0:7 .4 C q R1q ek = C q .1F qek = F q.1ek = 1 + 1:3q .5 shows the output and the sum of the square of the output for the regulators in a and b.1 .12 Introduce the polynomials A1q = AqDq B1q = B qDq C1q = AqC q and the noise e1 = e. 1:5 = .which yields the equations f1 . The output variance is E y2 = 1 + 1:32 2 = 2:69 2 d. : uk = .

deg B = d.3 is equivalent to qd.11.1 q q uk = .1 y k . A1 = AD and C1 = AC will also bo monic. BGDAFq yk = . deg B1 = deg A . The accumulated loss functions are shown in the lower diagram. 1. deeadbeat dashed and minimum variance solid. Let d1 = deg A1 .2 12.1 AqC q = Aq DqF1q + G1 q which implies that A must divide G1. D q q where F and G satisfy the equation q d.1 A . i.1C q = DqF q + Gq with deg F = d . B qG 1q q . 112 .3 where qd1 .4 100 Time 150 200 Since A.e.1 is then calculated as uk = . Figure 12. The minimum-variance controller for the rewritten system 12.1 C1q = A1qF1q + G1q Equation 12.10 Output 0 −10 10 0 50 100 150 200 Output 0 −10 500 0 50 100 150 200 Loss 0 0 50 The output and the sum of the square of the output for the two regulators in Problem 12.1qF q. Putting F = F1 gives the minimum-variance control law . We see that if A = D the controller reduces to the normal minimum variance controller. B G1qq y k q F 1 1 12. 1 and deg G = deg D . The deadbeat controller is used in the upper diagram and the minimum variance controller in the middle diagram. C and D are monic. G1q = Aq Gq .

1 ek : 1 .15 We have . 0::5q . ac yk .1 q y2 k = 1 + q. b yk . 017q.1 y1k = 1 .1 y1k .13 In this case Aq = q + a B q = b C q = q + c D q = q d=1 The identity in Problem 12. 1 + 1 .1 uk .12 gives q + c = q 1 + g0 The minimum variance controller is thus + c uk = . 1 b g0 = c Problem 12. 0 7q . cqbq a y k = .Problem 12.

1 = 0:94 .1 + q .1 uk . To normalize the notations it is convenient to introduce a new noise "k = ek .1 Gq .0:2y1k b. 1.1 113 . 0:7q . 0:5q . The minimum variance controller for y1 is then uk = . Assume that y1 can be measured. a.1 ek .1 1 = 1 + q .1 1 .1 This gives F q . 2 2:78 c.2 g0 + g1q .1 1 q 0:7q .1 = 1 + q . 1 = A B uk . The variances of y1 and y2 are 2 Ey1 = 1 1 2 Ey2 = 1 . 0:7q . 1 A 1 A .1 = 1 + q . 2 + 1 + 1.1 1 . 0:56q. 0:5 .11 + f1 q . 2 + C ek . The minimum variance controller for y2 when y2 is measurable is obtained by using the identity 1 . 1 q .

1 C G .1 y2k = 1 + q. Introduce the identity C = AA + q. It is now only necessary to predict one step ahead. 0q:.1 q y2 k + This gives .1G 1 1 This gives y2 k + 2 = ACA ek + 2 + ABA uk 1 1 G ek + 1 + B uk 1 = ek + 2 + A A A A 1 1 But This gives A "k + 1 = C y1k . B uk .and the controller 56 uk = .11 + q .1 "k = "k .1 ek + 1 The variances of the two signals are 2 Ey1 = 1 + + 12 + 2 Ey2 = 1 + 1 = 2 2 = 1:68 and d. 1 + ek . 0:941. In this case both y1 and y2 are measurable and y1 k will contain more recent information about the noise process than y2 k. 2 y1 k = 1 + q.

C . G q. B uk .1 uk 1 = "k + 2 + A A 1 1 AA C 1. 1 + B uk 1 y2 k + 2 = "k + 2 + A A C 1 C A A 1 1 G y k + B . A y k .

8 114 .1 y1 k . AGB y1 k 1 which is an admissible control law. For the speci c system we get .1 uk = . 00::56qq. 11. 1 1 1 = "k + 2 + C G y1 k + B uk A1 The variance is thus minimized when 1 uk = .

1uk . and Problem 12.1 yk = B q .1ek The variances of the two output signals are 2 Ey1 = 1 + 2 = 1:64 2 Ey2 = 1 We can thus conclude that the extra measurement will greatly improve the control of the output of the system.e.With this control law we get y2 k = "k = ek . that the process can be written as Aq.16 The same arguments can be used in this case as when deriving the normal minimum variance controller.17 can be used here also and gives yk + d = C ek + d + B uk + D vk A A A ek + d + G ek + B uk + D v k =F A A A G .1 y2k + 1 = 1 + q.1v k . Also assume that deg A = deg B = deg D i. 1 and y1k = 1 + q.1"k + 1 = 1 + q . d + C q.1ek + D q . Assume that the system has a stable inverse. d The identity 12.

d . BGF y k . D v k B 115 . q. q . D vk .d G v k 1 = F ek + d + C Gy k + B F uk + D F v k The minimum variance controller is thus uk = .d G uk C AC + ADC C . B uk . A y k . d = F ek + d + A C C C + B uk + D v k A A ek + d + G y k + B =F C .

0:9 + 1 1 r1 + p2 + rp1 z + rp1 z . 0:9 z .1 rz + p1z . q s0 qr y k + .5. 0:9.1 P z contains stable zeros of the right hand expression Lemma 12. C q = q . Let P z be the closed loop system characteristic equation 12. The Diophantine equation to be solved is S uk = .1 + B z Bz. Bq = q.17 Aqyk = Bquk + C q ek Aq = q .0:9 + r1 + s0 . Rq y k q PC = AR + BS or This gives z + p1 z .0:9 r1 + p2 = 1 + 1:81 1 which has two solutions.1 . 0:9 z 1 This gives the system of equations rp1 = . 0:5 = z . 1 + :1:81 + 1 + :1:81 2 .1 .0:5p1 = . 0:9z.0:9r1 s0 = 0:4 + 4 p1 9 1 r1 = 5 p1 9 which results in the controller 116 uk = .45 rP zP z. one of which gives a stable P z p1 = . 0:5 Control law: r where S 0 = 0. 0:9z + r1 + zs0 The solution is given by p1 .1 + p1 = z . 1 18 18 Determine the LQG-regulator by means of pole placement: Am = P z = z + p1 Ao = C z = z .1 = 1:81 + 1 . See computational procedure in Section 12. 0:5 = . 0:5 LQG-Control: Minimize E y2 + u2.1 = AzAz.Problem 12.

The input u is For the closed loop system we get B uk + C ek = B .

1 B q = h C q = q + c The minimum variance regulator is given by the Diophantine PC = AR + BS where The solution is P q = q d. R y k = .0:36 1:030 10 . S yk + C ek y k = A A A R A or CR q+r yk = ARCRBS ek = PC ek = R ek = q + p1 ek + P 1 Theorem 10. P ek = .0p2 1 p1 V ar y 0:1 .0:70 1:197 V ar u 0:135 0:0658 0:0148 Problem 12.16 we get a polynomial description of the process Aq y k = B q uk + C qek where Aq = q .4 gives In the following table the calculations are summarized for = 0:1.1 Bq = h r0 = h s0 = c + 1 and the minimum variance regulator is uk = . V ar y = 1 + r1.24 From Example 12. p2r1p1 1 2 1 S R S S uk = .4 gives 2. R P ek = .. s2 V ar u = 1 . 1 and 10. c + 1 y k h 117 .0:077 1:0012 1 . q s0 qp ek + 1 Theorem 10.

8 we get Q1 = Q12 = Zh Zh 0 1 ds = h 2 and thus The Riccati equation is S = T S + Q1 .. 2 . h h 2+ 3 and P z = z + p1 To get the regulator we solve the Diophantine see p. design we use a state-space description of the process see Example 12.T S .T S + Q12 and the solution is S = ph 12 p 1 L = h 3+p3 2+ 3 The closed loop system has a pole in .T S . From 11. q .p1 = .L L = Q2 + . 0:2 C z = z 2 + 0:4z 118 .p1 p 1 3 + p3 = p3 . 1:4z + 0:65 B z = z .p1c 1 s0 = h p1 + c + 1 + p1 c 1 h p1 + c + 1 + p1cq y k uk = .1.L = 1 . 481 PC = AR + BS S 0 = 0 q + p1 q + c = q . p1 c s ds = h 2 Z0 h 3 Q2 = s2 ds = h 3 0 Problem 12.29 The process is described by Az = z2 . 1q + r1 + hs0 q which has the solution r0 = . . v kh ykh = xkh + kh To obtain the LQ-controller we have to sample the loss function.6 .11.16 xkh + h = xkh + hukh + v kh + h .In LQ . LT Q2 + .

0:65 . BF y = ay The closed loop system becomes y k = ek + cek . The dead-beat controller is obtained from the identity AzF z + Gz = z 2 Gz = 1:4z . : u = . 00265 y q : c.a G u = . 1 and the variance vary = 1 + c2 119 . Minimum variance control vary = 1 Dead-beat control vary = 1 + 0:42 2 = 4:64 2=4 which gives Problem 12. To get the minimum variance controller we use the identity Az F z + Gz = z d. Assume that C = 0 and use the identity C = AF + G This gives and the controller F =1 G = .30 a. 1:4q. 1:4z + 0:65 + g0z + g1 = z 2 + 0:4z This gives The control law is Gz = 1:8z . 0:65 G . 1:8q.a.1 C z z 2 . BF y = . : u = . 00265 y q : b.

1 . Replace C by a new polynomial obtained from spectral factorization C z C z . 1.a ^ The closed loop system is now q c yk = q + c ek +^ which has the variance. 1:25z .b. see Theorem 6. 0:8z . 1:25 = 1:252z .10:+q0:3 "k :q z m. 0:09 yk + 2jk = qGqq y k = 0:03q0.31 The C polynomial has a pole outside the unit circle. : E y = 1:2521 + 0:32 = 1:70 ~ b.4 ^ vary = 1 +1c. 2cc 2 ^ 2 It is better to use a guessed value c if ^ 2 0 c 1 +cc2 ^ Problem 12.83 y k ^ C q. The minimum variance controller for this ^ model is given by F =1 G = c.1 = z . 0:8 The new process is where " has the standard deviation 1:25. The 2-step ahead predictor is given by q2 8 yk = q2 . a.1 C z = Az F z + Gz zz2 . 0:8z = z2 . Assume that C z = z + c.1 . c. The prediction error variance is 120 . 1:1z + 0:3z + f1 + g0z + g1 which gives and F z = z + 0:3 Gz = 0:03z .

z = . 1:7z2 + 0:8z . 0:1f0q 2 + f1 q + f2 + .z with B . 1:18z + 0:16 and the controller is G 92 2 + :16 uk = .0:1f0 + 0:8f1 . 0:9 =q3 . q monic. q =Aq F q + B .0:9q + 1g0q 2 + g1 q + g2 This gives the system of equations 1 = f0 .1 C q B .2 B. Since B z has a root outside the unit circle we use the procedure in Theorem 12. z3 z . so B + z = . 2g1 + 1:8g2 0 = . BF y k = . is given by vary = 1 + f12 = 1 + 1:62 = 3:56 c.32 a.1:862 g2 = 0:272 121 . 0:1 = z 3 .1 = . 2g2 which has the solution f0 = 1:000 f1 = 0:700 f2 = 2:721 g0 = 2:490 g1 = . 0:1q . 2g0 + 1:8g1 0 = . 1:7f1 + f2 + 1:8g0 0 = . This gives the identity zC z = AzF z + Gz where deg F = 1 and deg G = 2. d = 2.0:1f2 .e. 1:7f2 .Problem 12.3.. 1:2qq.1:7f0 + f1 0:09 = 0:8f0 . The output variance when using the minimum variance controller in a..0:1f1 + 0:8f2 . zGq q3q . 0:1z + f1 + g0z 2 + g1z + g2 The solution is F z = z + 1:6 Gz = 1:92z 2 . 1:7q2 + 0:8q .:1:18q+ 106 y k 0 9q : b. i. Factor B z as Bz = B + z B.0:9q + 1 The Diophantine to solve is qd.

1:245q 0:70q:931q:7210:136 y k 2+ q +2 The closed loop system is .31 gives 2+ . uk = . B +GqqF q yk = .Equation 12. .

33 Determine the controller from q q. ek 1 + GB 2 + 0:7q + 2:721 = q q q . B + F y k + Cek = . 0:7q + r1 + 0:9q + 1s0q + s1 = q 2q .0:368 yk = ARCRBS ek = CR = 1 + 0:526q.0:7 0:7 .1 C q . : :9q yk = q . G Ayk = Buk + Cek = B . ek = qd. GB y k + Cek F .4 gives the output variance V ar y = 94:7 Problem 12. 1:7r1 + 0:9s1 + s0 = 0 0:7r1 + s1 = 0 r1 = 0:526 s0 = 0:526 s1 = .9 p. ek = qd.FB.01q+ 10:7 uk + q . 468 var y k = 20 2 = 1:053 2 19 122 with the solution . 0 AR + BS = qd. 0:7 This gives the system of equations . 0:9 ek Theorem 10.1 CB .1:7 + r1 + 0:9s0 = .1q0.7 :7 ek . or CF yk = AF CF . 1q .1ek + qC var y k = 1 + 0:5262 2 = 1:27 2 Compare to Example 12.

- ASTROM - Computer Controlled Systems (3rd Edition) Parte I
- Soluções do Livro de Controle Digital
- Adaptive Control 2nd. Edt. by Karl.J.astrom - Solution Manuel
- Computer Controlled Systems - Theory and Design
- Digital Control of Dynamic Systems
- Digital Control of Dynamic Systems (1998)
- Linear State Space Control Systems
- Discrete-Time Signals and Systems
- PSIM Tutorial
- Preface.doc
- Lab 1
- Nagoor Kani Control Systems
- Design of a PID controller using Matlab and LabVIEW
- Slides Fall 10
- Practical Issues in Model-based Surge Control for Centrifugal
- 198209370-EE2257-SET1
- Differential Equations - Solved Assignments - Semester Summer 2007
- Classical Control Systems Book
- 126U2
- LAMI MAIN.pdf
- Study Problems
- Math Paper 3 Practice
- EEE_402_exp1
- Add Math Comman Mistakes - TEACHER
- IIT JEE Mathematics 1997
- Solutions 3
- Definitions and Basics TRIGNOMETRY

Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

We've moved you to where you read on your other device.

Get the full title to continue

Get the full title to continue reading from where you left off, or restart the preview.

scribd