Զѧ
ACTA AUTOMATICA SINICA
1998 24 5  Vol.24  No.5 1998



µƲ

ժҪоĺƲϵͳһоĻĿҪźĳЩضб仯ʱϵͳӦƲڵʱ̬֮̵̬ĵʱӦܵĶ.Ϊϵͳ³ԣҪƷϵͳʱϵͳĵʱӳ.⣬ΪһϵͳԶԣҪ.Ƕ޵ʱϵͳԣʱϵͳѲ޲ϵͳ.̽ĳЩ֮ϵͳµƲ⣬ͬʱʹMATLABмƲ.
ؼʡƲϵͳ룬ʵʽ.
THE RIPPLE-FREE TRACKING PROBLEM
FOR ADMISSIBLE INPUT
CHEN XIAOTIAN
(Dept.of Electronics and Information Engineering,Guangdong University of
Technology Guangdong,Guangzhou510643)
AbstractThe design of minimum settling-time and ripple-free digital control systems are investigated in this paper.The basic specifications for the system are that with a reference signal within a pre-specified class the system's response can have no ripple and no error after the settling-time,and the settling-time for the transient response must be as short as possible.To improve robustness and flexibility to input,it is necessary to design the non-minimum settlingtime system and introduce a staleness-weighting factor.This is a compromise design between the limited settling-time and the overshoot.Some fundamental algebraic concepts are introduced.Then the ripple-free tracking problem for admissible input and its CAD with MATLAB are considered.Simulative examples are reported.
Key wordsMinimum settling-time and ripple-free digital control systems,admissible input,algebra realized condition,polynomials.
1
ĿǰʹּϵͳѾ㷺رܣݿϵͳʱڲܾϵͳͬ,ֿϵͳ޸ںƳϵͳ.ź,ĺ޲ϵͳֻڲϲ壬ڲ֮,һ.ڱصϵͳһ㶼,пڲ֮,ԲܹܳӦĻек,ҪϵͳڲоȷĹ켣ԵرҪ.۵ݣ漰뼯ԼƲֿϵͳĻƷ.
һЩ˶ĿƵй㷺о,TouKuoOgata(1987)Jetto(1989) һҪ.ʱ״̬Kalman˻.ҲBhattachar-yyaо.ǶӲͬǶЩ.Ŀǰ˸ȤźĳЩض͵Ʋ.ʵ,ڶֹҵӦó,ҪϵͳڰĳЩض͵źŸλ.,ͳƷԵòЧ,һܲһЩϵ.Ϊ,бҪһǶ̽һ.Ϊ˱Ͻķ,һЩ.
2
1) Fp(d)ΪFϱԪΪd=z-1Ķʽ.
2) a(d)Fp(d),a(d)dݳΪa(d)Ľ״,Ϊa.a(d)δ֪ϵĸΪa+1.
3) a(d)Fp(d)a(0)=1,a(d)Ϊһʽ.һʽļFcp(d)ʾ,
Fcp(d)={a(d)a(d)Fp(d),a(0)=1},
a(d)δ֪ϵĸΪa.
4) ɷֱʽa(d)b(d)ĶʽΪʽСʽ,Ϊa(d),b(d).ɷֱnʽa1(d),a2(d)ai(d),(i=1,2,n)ĶʽΪnʽСʽ,Ϊai(d).
5) F(d)ΪFϱԪΪd=z-1ʽ.
6) ,c(d)Ϊ̬ʽ,̬ʽļ(d)ʾ.
7) c(d)(d){cn}{cn,cn+1,},c(d)Ϊȶĳ̬ʽ,ȶĳ̬ʽļF+(d)ʾ.
8) a(d)p(d)1/a(d)+(d),a(d)Ϊȶʽ.
9) a(d)Fcp(d)ȶʽӼΪa+(d)Fcp(d),ʽӼΪa-(d)Fcp(d),a(d)=a+(d)a-(d).
3Ĳ
Ʋֿϵͳͼͼʾ123:
һֿG(d)(d)вĵλϵͳ.бԪd=z-1.U(d)(d)ΪĳЩض޼


ͼ1Ʋѧϵͳͼ
(1)
ri(d)Fp(d),vi(d)Fcp(d).
E(d),M(d)Y(d)(d)ֱʾ,źŵļ,We(d),Wm(d),W(d)ʾӦ,ͱջ崫ݺ.P(d)=b(d)/a(d)(d),a(d)Fcp(d),b(d)Fp(d),b(0)=0Ϊɢض崫ݺ.öɲ(һΪױ)ضP(s),ٰһd=z-1ӣ45.
ҪһG(d),ʹϵͳ޸֮,ڼU(d)ÿһ,ڲ̬,ضƲ.ϵͳ
1)Ҫ:뼯E(d)Ӧпܵ{e(i)n}nieʧ,Ei(d)Fp(d)
2)ƲҪ:ΪʹضƲ,Ӱضп{m(i)n}ĺﵽȶ״̬,񵴵Ŀ崫ݺWm(d)Fp(d)ӦΪ޶ʽ.
4Ľ
ݿͼ,ǵСλ,
(2)
뱻ضP(d)ֲȶ,崫ݺWe(d)ضĸa(d)вȶa-(d).ٿǵҪ󼰱ضٰһd=z-1,
We(d)=c(d)v(d)Fcp(d),(3)

v(d)=a-(d),vi(d)ݣ=a-(d)q(d)Fcp(d).(4)
c(d)q(d)Fcp(d),c(d)ΪһĴʽ.ʹ
Ei(d)=Ui(d)We(d)Fp(d)(5)
Ϊ޶ʽ.ٸƲҪ,
Wm(d)=s(d)a(d),(6)
Сs(d)Fp(d)Ϊʽ.
ʽ(3)(6)ȷ˱ջ崫ݺ
W(d)=1-We(d)=Wm(d)P(d)=s(d)b(d)Fp(d)(7)
Ϊd=z-1ӵ޶ʽ.
ʽͨȽϵ,ʽ(3)ʽ(7)
s(d)b(d)=1-c(d)v(d).(8)
ʽδ֪ϵΪc+s+1,ڵʽ˾޳,ڱȽͬϵķ̸Ϊs+b=c+v.ͽ׽,
s=v-1c=b-1.(9)(10)
νĴʵARC(Algebraic Realizability Conditions).Ϊһϵͳܶ,ʽĽ״λӦӣ45.,Ϊʹ÷н,b(d)v(d).õʵֵĿ
(11)
(1)(3)(5)(10)ʽ,ӦСʱΪ
nie=ri-vi+b-1+v,(12)
5
ƲĽƷĳЩֿϵͳƻ5.ֻϵͳĺƲ,Ҳȶ㼫ķ,Ƴϵͳ,³Զ̫,ʵǲֱʹõ.бҪһ̷,ʹƽ߱ӦеĹ.ָĹʵʶѧģ͵Ĳб仯ʱ,ƲӦδ?ϵͳֹ޷ܵĿ򳬵ʱ,ƲӦδȵ.
,һ㶼ԲϵͳԵİ취Խ,ҲǽϵͳƳɷϵͳ.취,Ҫϵͳ崫ݺWe(d)W(d)Ľ״,Ҳ(910)ʽĻӦcs.ÿһ״λͬʱӴʽδ֪ϵ,ԿɽһЩҪԼ,ΪٳȽĳЩؼδ֪ϵȷΪ֪ϵ,ΪٿÿƳֵm(0),m(1)ȵ.Ȼʱϵͳĵʱӳ,ϵͳ,³ԵȶԵõ.ϵͳԲжȽ߶.ʵʶѧģ͵Ĳб仯ʱ,ϵͳܱ.ע⵽,ܵĸΪ۵,ᵽķ,֮еĽܱϵͳΪƲϵͳ.
йԳӰԽĶԷ.οDahlin㷨ȵ˼·,(1)(5)ʽ,We(d)Fcp(d),vi(d)Fcp(d),ri(d)Fp(d),Ei(d)пܾдʱ.We(d)һϵԽ,۳ʱһԽ,⽫ϵͳܴӰ.޶ʽWe(d)һϵĳ̶ֳϷӳϵͳֳܵˮƽ.(4)ʽ, vi(d)We(d),ԵUi(d)U(d), Uj(d)U(d),ijvj(d)ֲvi(d)ʱ, vj(d)һϵĳ̶ֳϷӳUi(d)ϵͳʱĳ.
ҪС,ɳᵽӴʽ״εİ취.ӽ״ζɶ,ʽ,ΪWe(d)ӵc(d)һϵȷΪһԼСֵ.
ʱԲ㹤ϵҪ,ɽһùӷ,ֱӸWe(d)ZƽҰһȶ.Ȼȫ޲Ϊ۵.ȡ˶нӦЧ.ʵֹӷиֲͬ;,һ.ƷһصǿԭƻֱӽϿԽܵˮƽ.ΪZƽҰӵȶ
(13)
Ȼ,ɸ(2)ʽֱWe(d)We(d)ϵͳƳ.
6ʵ
ֿϵͳ1
(14)
οΪ
{u(1)n}=exp(-0.2n),(15)

{u(2)n}=sin(/4)n,(16)

(17)

(18)
ָΪi)ϵͳڵʱ֮ӦΪƲ, ii)̬ӦĵʱӦܵض,ֿЩҪ.
a(d)=a+(d)=(1-0.368d)(1-0.135d),(19)
a-(d)=1,b(d)=0.399d+0.147d2(20),(21)
r1(d)=1,v1(d)=1-0.819d,(22)
r2(d)=0.707d, v2(d)=1-1.414d+d2.(23)
ݹʽ(4)
v(d)=q(d)=(1-0.819d)(1-1.414d+d2)(24)

r1=0,v1=1,r2=1,v2=2,v=3,b=2.(25)
ݹʽ(9)(10)
s=v-1=2c=b-1=1.(26)(27)

s(d)=s0+s1d+s2d2,c(d)=1+c1d.(28),(29)
ٸݹʽ(8)
(s0+s1d+s2d2)(0.399d+0.147d2)=
1-(1+c1d)(1-0.819d)(1-1.414d+d2).(30)
ΪȽͬϵ϶ķ,ʹMATLABԽм.MATLABԷе:
s(d)=4.6966-5.1296d+2.0005d2,(31)
c(d)=1+0.3591d.(32)
빫ʽ(11)
(33)
ʽ(12)֪,СʱΪ
n1e=r1-v1+b-1+v=3,n2e=r2-v2+b-1+v=3(34),(35)
֮,Ϊ,MATLABнȷǺ.ǰ, 
E1(d)=c(d)v(d)U1(d)=(1+0.3591d)(1-1.414d+d2)=
1-1.0549d+0.4923d2+0.3591d3.(36)
еڶϵΪ-1.054 9,˵100%,ϵͳǲʵʩ,Աѳѹ.ݽܵĴķ,s(d)c(d)Ľ״βؼc(d)һϵΪ1,
s(d)=s0+s1d+s2d2+s3d3,
(37)
c(d)=1+d+c2d2(38)
ٸʽ(8)
(s0+s1d+s2d2+s3d3)(0.399d+0.147d2)=
1-(1+d+c2d2)(1-0.819d)(1-1.414d+d2).(39)
ȻʹMATLABԳȽͬϵ,Ϊ
s(d)=3.090 2-1.542 5d-1.466 2d2+1.315 6d3(40)
c(d)=1+d+0.236 1d2.(41)
ʽ(11)
(42)
ʱ
E1(d)=c(d)v(d)U1(d)=(1+d+0.236 1d2)(1-1.414d+d2)=
1-0.414d-0.177 9d2+0.666 1d3+0.236 1d4.(43)
ʱѹ42%,Ȼϼֻ˲ϵֵ,֤1,t5T,tR(TΪ), ʵʷֱexp(-0.2t)sin(/4)t,ʵƲ.
ֻǾС,Եʱָ֮ûرҪ,Բùӷ.ӷԲMATLABԳ,Ϊ˼,ֻ0.1,0.50.9֮e1e2.1
1


ne1e2
0.10.50.90.10.50.9
01.1.1.000
1-0.3140.0860.4860.7070.7070.707
2-0.209 3-0.134 90.259 50.198 70.481 50.764 3
30.645 20.598 70.899 7-0.392 2-0.171 40.275 8
40.300 70.535 51.045 9-0.176-0.222 40.111 4
50.030 10.267 70.941 3-0.017 6-0.111 20.100 3
60.0030.133 90.847 1-0.001 8-0.055 60.090 3
70.000 30.066 90.762 4-0.000 2-0.027 80.081 2
80.033 50.686 2-0.013 90.073 1
90.016 70.617 6-0.0070.065 8
100.008 40.555 8-0.003 50.059 2
110.004 20.500 2-0.001 70.053 3
120.002 10.450 2-0.000 90.048
130.0010.405 2-0.000 40.043 2
140.000 50.364 7-0.000 20.038 9
150.000 30.328 2-0.000 10.035
160.000 10.295 4-0.000 10.031 5
170.000 10.265 80.028 3
180.239 30.025 5
190.215 30.022 9

ӷеԿ, мֵԽ,̬Խƽ,ʱҲԼӳ.Ҫùӷ,µĹǸݾҪѡ񼫵,Ȼ(13)(2)ʽ.
7С
֤˵Ϊضʱ,ƲֿȷʽĽʽ,ƲֿϵͳĽƷĳЩֿϵͳֱƵһ.,ֿG(d)崫ݺɹʽ

,вֿ֪ͨ롢ضԼʽ
a(d)=a+(d)a-(d),v(d)=a-(d),vi(d)ݣ=a-(d)q(d),
s(d)b(d)=1-c(d)v(d)
,b(d)v(d)ʱ,ͽ׽ĴʵΪ
s=v-1,c=b-1.
߼飺졡1945.Ϊ㶫ҵѧϢϵϢ̽Ҹڡ.1968ҵͨѧߵ繤ϵ, 1978꿼йƼѧߵѧϵУʦʰѧϰ, 19901991ӢBradfordѧϵоԱ.ĿǰҪоΪơֵ㷽ݴϵͳ漰źŴ.
ߵλ(㶫ҵѧϢϵݡ510643)
ο
1Jetto L. Ripple-free tracking problem. Int.J.Control.1989,50:349359
2Ogata K. Discrete-Time Control System. Englewood Cliffs,NJ:Prentice-Hall,1987
3Wang S S,Chen B S.Simultaneous dead-bet tracking controller synthesis.Int.J.Control.1986,44:1579
4.The Order of the Unspecified Polynomial in Ripple-free Tracking Problem,ܿԶ.:ѧ,1993.25742583
5Chen Xiaotian,Li Zhang,Readle J C.The analytic formulas of the ripple-free tracking problem. Automatica.1994,30(5):905908
ոڡ1996-06-20
