Զѧ
ACTA AUTOMATICA SINICA
1997 23 1 Vol.23 No.1 1997



һֿƻеֵԵģ߼
̷Ⱥΰǿ
ժҪһֻеֵģ߼Ƶ·. ܸϵͳǰӦԶ޸仯ʵĿ. оԿӰĻϣȷһ. ȻNelder-Mead㷨Żģ߼. øģһзԶѧɶȻе֣֤᷽ЧԺ³. 
ؼʡģƣѧŻԿ.
A SELF-TUNING FUZZY CONTROLLER FOR MANIPULATORS
TAN QUNHUA LI WEI WANG YONGQIANG SUN ZENGQI
(Dept. of Computer Science & Technology, Tsinghua Univ. Beijing 100084)
Abstract This paper presents a new approach to automatic tuning of fuzzy logic controller for manipulators, which modifies the shape of membership functions automatically according to previous responses to achieve a desired control performance. On the basis of defining a set of cubic splines as membership functions regarding change-in-error j(t), the optimal membership functions are found by the Nelder and Mead's simplex algorithm.
The proposed control scheme consists of a traditional fuzzy logic controller and a conventional derivative controller. In order to demonstrate the effectiveness and the robustness of the control scheme, we report a number of simulation results of a two-link manipulator in the presence of load and friction changes.
Key wordsFuzzy logic control, robotics, optimum technique, nonlinear control.
1
ڹȥĶʮУģ߼ѾɹӦзԶѧ̻δ֪ṹϵͳΪֿģģ˵ľ֪ʶɵ. Ȼģ߼Żȱ. 
Żģ߼ķ֮һǸϵͳǰӦ֯㷨Զ޸ĹԴﵽĿЧ1-3. ʵ޸ĹǻһָŻһľ߱. ˣҪڽ϶̲ʱɹŻһĹ. 
ͼ1ʾĿƽṹһԶģ·ƻе[4-6].Ҫ˼Զ޸״ǹ. Ϊˣȶһ仯(t)ȻϵͳǰӦNelder-Mead㷨Ż. ÷ȴͳ֯ΪЧΪŻ߱תΪѰҾĳЩض״Ĳʸ. ⣬ģҲױƹʦ. 


ͼ1ģĽṹͼ
2ģ߼Ʒ
ģ7Ŀƽṹһͳģһ΢ֿ(ͼ1). ģźǻеֽǹؽjֵʵֵej(t)=refj-j(t)Լı仯j(t)źuj(t)Min-Max㷨ģ1ʾ. ģģʵתģķ. ƷṩеֽǹؽڵĿئj(t)
(1)
1ģƹ

eNBNSZOPSPB

NBPBPBPBPSZO
NSPBPBPSNONS
ZOPBPSZONSNB
PSPSZONSNBNB
PBZONSNBNBNB

յĿj(t)ɣ 1) ģźuj(t)ıŴҪǵej(t)ܴʱϵͳӦ 2) ģźuj(t)Ļֲ Ҫڵj(t)ӽrefjʱӦľ̬3) ٶȷźj(t)΢ֲϵͳȶ. Kpj,KIjKDjZiegler-Nichols.
3ѡȡ
һ˵ģгʹλָε. ıЩ״Ӱ. Ϊ˻úõĿԣģ4Ƕ仯j(t)޸ģΪٶȷ. ޸ĵķǣ ԭʼ仯j(t)Ӳֵ㣬Ӷٱ޸ĵĸ. ģ8еоӦԵȲκҲָڱj(t)Ϊֵͼ2bʾ. Щ״ͨƶǵġ㡱ı. Щ㡱ͼ3иƶøıǵĲʵֵ. kb,kskzΪʸk=(kz,ks,kb)ҷֱ(NB),(PB),(NS),(PS),Լ(ZO)ͼ3a-cʾ. ʸkܹڣ0.15,0.85ݡ0.15,0.85ݡ0.15,0.85ݱ仯. ෴e(t)uj(t)ѡΪκϵͳйв䣬Ƿֱͼ2aͼ2cʾ.




ͼ2Ա



ͼ3ֵ
4Ż㷨
ķУԵǸϵͳǰӦͨʸѵK*=(k*z,k*s,k*b)õ. ƽ׼
(2)
Ӧ. ΪӦ仯j(t)Ĳʸk=(kz,ks,kb)ϵͳӦżӰ죬J=f(kz,ks,kb)J=f(k)ǲʸk=(kz,ks,kb)ĺ. Ϊ˻ĿԣǱͨӦʸk*=(k*z,k*s,k*b)Сֵf(k*). ΪϵͳԻ侫ȷĶѧ̣ԺѵõJ=f(k)Ľʽ. ΪNelder-Mead㷨Щȱģ͵Ż⣬Ż. ǵľ飬(ZO)ϵͳĿӰԲkzϵͳй̼Ϊ. Nelder-Mead㷨Сֵf(k*)£6ݣ
1) 㿪ʼk1(c,k(1)s,k(1)b),k2(c,k(2)s,k(2)b),k3(c,k(3)s,k(3)b)ҼJ1=f(k1),J2=f(k2)J3=f(k3).
2) ҳֵJh, мֵJgСֵJ1ԼӦĵkh,kgk1.
3) ҳkgk1е㣺 kf=0.5(kg+k1),ҼJf=f(kf).
4) kfķkh,kʺJr=f(kr).
5) JrJg,kc=kh+(1-)kfkcJc=f(kc)ѹ(0£1)ѹϵ. JrJgkc=kr+(1-)kfkcJc=f(kc).
6) JcJh,kckhԣصڶ. JcJhһ.
7) kh=kh+0.5(kh-k1)kg=kg+0.5(kg-k1)Сεĳߴ磬ҼJhJg. صڶ.
5еֵĿ
Ϊ֤᷽ЧԺ³ԣһзԶѧɶȻе֣ûеֵĶѧ̺Ͳģ8еõ. ĵڹ£ȣZiegler-Nichols ȷĲKPjKIjKDjʹеֻһɽܵĿӦȻڴ˻ϣNelder-MeadŻ㷨ģʹеֻһĿӦ. 
ͼ4a-bеиغĦ£ģPIDӦ. ǵĲ޸غĦµڳѿƲ. ȻиغĦʱPIDõӦԱ仵. ԭ޸غĦʱȷPIDƲʺϿиغĦĲȷԱ仯ϵͳ. ȻģиغĦʱеֵĿȻܺ(ͼ4(b)).


(a) ؽ1(PID)(b) ؽ1(ģ)
ͼ4PIDƽģƽ
ķھзԶѧ̼ģͲȷϵͳ.
ʽĿȻѧ𡱺͡ҽίʦĿ.
ߵλ廪ѧѧ뼼ϵ100084
ο
[1]Procyk T J, Mamdani E H. A linguistic self-organizing process controller. Automatica, 1979, 1515-30.
[2]Scharf E M, Mandic N J. The application of a fuzzy controller to the control of a multi-degree-freedom robot arm. Industrial Application of Fuzzy Control, 1985, 41-61.
[3]Wakileh B A M, Gill K F. Robot control using self-organizing fuzzy logic. Computers in Industry, 1990, 15175-186. 
[4] Li W. A new self-organizing fuzzy controller. The 3rd International Workshop on Advanced Motiion Control, at the University of California at Berkeley, 1994.
[5]Li W, Sun Z Q, Janoscha H. An approach to automatic tuning of a fuzzy logic controller for manipulators. Proc. of IEEE/RSJ International Conference On Intelligent Robots and Systems, 1994. 
[6]Li W. Optimazition of a fuzzy logic controller using neural network. IEEE World Congress On Computational Intelligence, Proc. of FUZZIEEE '94, 1994 1223-227.
[7]Li W et al. Fuzzy control of robotic manipulators in the presence of joint friction and loads changes. The 1993 ASME International Computers in Engineering Conference, San Diego, USA, 1993.
[8]Li W, Wu Z W. A self-organizing fuzzy controller using neural networks. The 1994 ASME International Computers in Engineering Conference, 1994, 2807-812.
ոڡ1994-10-13
