王策 楊升 張磊 左信 白小波
摘 要 機械臂系統(tǒng)具有非線性和不確定性,針對其精確跟蹤控制問題,提出基于指數(shù)增益迭代學習的二階滑??刂扑惴āJ紫?,采用滑模控制增強系統(tǒng)對外界非重復干擾的魯棒性,采用二階滑模方法使控制器產(chǎn)生連續(xù)的控制信號,以減輕傳統(tǒng)滑??刂频亩墩駟栴};其次,結合指數(shù)增益迭代學習控制逼近由摩擦力等周期性擾動引起的有界重復性擾動,通過設置指數(shù)增益系數(shù)以提高迭代的收斂速度,該方法既能減小系統(tǒng)的跟蹤誤差又能保證系統(tǒng)的全局漸近穩(wěn)定性,提高系統(tǒng)的動態(tài)性能;最后,分析算法收斂的充分條件,以二自由度機械臂為例對所提方法進行理論數(shù)值仿真。仿真結果表明:所提控制方法可以保證系統(tǒng)的穩(wěn)定運行,輸出跟蹤誤差收斂到接近零的鄰域,提高了系統(tǒng)的控制精度,對干擾保持了良好的魯棒性。
關鍵詞 二階滑??刂?指數(shù)增益迭代學習控制 機械臂系統(tǒng) 軌跡跟蹤 抖振 控制精度
中圖分類號 TP18? ?文獻標志碼 A? ?文章編號 1000-3932(2023)05-0644-08
機械臂在工業(yè)生產(chǎn)中廣泛應用,越來越受到工業(yè)界和學術界的重視,軌跡跟蹤控制是機械臂的重點研究方向之一。機械臂軌跡跟蹤控制的主要目標是通過給定各關節(jié)的驅動力矩,使得機械臂的位置、速度等狀態(tài)變量跟蹤給定的理想軌跡。在實際控制工程中,難以用精確的數(shù)學模型描述其動態(tài)特性,因為機械臂系統(tǒng)是一個具有非線性、不確定性的系統(tǒng),存在很多不利于提高系統(tǒng)性能的因素,如:非線性因素,包括電機力矩波動、驅動飽和、耦合力矩、干擾力矩等;參數(shù)變化,包括負載變化導致的轉動慣量變化、溫度升高導致的參數(shù)變化等;測量延遲及測量噪聲等因素。因此,在建立數(shù)學模型時需要對不確定因素作合理的近似處理,然而由近似模型來設計控制器,設計中被忽略的不確定因素往往會影響控制系統(tǒng)的品質,甚至導致系統(tǒng)不穩(wěn)定[1]。
滑??刂疲?~4]提出通過切換控制變量使系統(tǒng)按照預定的滑動模態(tài)的狀態(tài)軌跡運動,保證被控系統(tǒng)在參數(shù)攝動和外界擾動情況下的穩(wěn)定性,具有較強的魯棒性,在機器人控制中得到了廣泛的應用。文獻[5]針對參數(shù)未知和外界干擾條件下的移動機械臂,提出了一種基于模糊小波神經(jīng)網(wǎng)絡的自適應分數(shù)階非奇異終端滑??刂疲抡鎸嶒烌炞C了方法的可行性和有效性;文獻[6]設計了滑模控制器調節(jié)機械臂的死區(qū)和重力未知行為。然而,傳統(tǒng)滑??刂戚敵龅牟贿B續(xù)切換會引起抖振效應,在實際操作中,控制信號的不連續(xù)會損壞執(zhí)行器或控制裝置。邊界層法和飽和函數(shù)法是解決抖振問題的常用方法[7,8]。但是,飽和函數(shù)替代符號函數(shù)法在減小抖振和保證控制精度之間難以兩全,高階滑模算法是一種利用滑模積分項有效減輕抖振的方法[9,10]。文獻[11,12]提出的高階滑模理論被認為是減小滑模抖振且保持系統(tǒng)較高控制精度的有效手段,并提出幾種二階滑??刂扑惴ā6A滑模的核心思想是將不連續(xù)切換項轉移到滑模面的二階導數(shù)上,進而通過積分作用實現(xiàn)控制輸出的連續(xù)性,因此可以有效抑制抖振,同時保持魯棒性。文獻[13]提出一種二階滑??刂扑惴ǎ瑧糜谝活惗噍斎攵噍敵鱿到y(tǒng),該算法不需要滑動曲面的導數(shù),所得到的控制是動態(tài)的,減輕了系統(tǒng)輸入的抖振。盡管滑??刂扑惴ê唵?、魯棒性強,但實際機械臂控制系統(tǒng)中模型不確定性的存在使得僅使用該算法時跟蹤效果不夠理想。
迭代學習控制(ILC)的優(yōu)勢以其在控制具有部分未知和時變參數(shù)的重復系統(tǒng)方面的性能得以體現(xiàn),并在控制領域得到廣泛運用[14]。基本的ILC方法是根據(jù)過去迭代得到的跟蹤或性能誤差,在下次系統(tǒng)運行時調整控制輸入,從而改善瞬態(tài)響應,最終達到預期軌跡[15~17]。文獻[18]設計了一種迭代學習控制器,以選擇適當?shù)膶W習增益實現(xiàn)任意高精度的輸出跟蹤,而不考慮測量誤差。文獻[19]對具有初始狀態(tài)誤差的p型迭代學習控制的收斂性進行分析,仿真結果證明了算法的有效性。文獻[20,21]提出通過將滑??刂谱鳛轸敯舨糠趾蛯W習控制作為智能部分相結合的控制,既具有滑??刂频聂敯粜杂志哂械鷮W習的高精度。文獻[22]提出一種具有連續(xù)滑模的魯棒迭代學習策略,并在SRV02旋轉裝置中進行驗證,實驗結果表明該算法的控制信號是連續(xù)的,具有良好的輸出跟蹤性能。
筆者提出一種基于指數(shù)增益迭代學習的二階滑??刂扑惴?,首先利用二階滑??刂铺岣呦到y(tǒng)對外界非重復性干擾的魯棒性,二階滑模方法用以減輕抖振效應,然后采用指數(shù)增益迭代學習控制對系統(tǒng)模型和外界重復性擾動進行逼近,擺脫對特定數(shù)學模型的依賴,通過設置指數(shù)增益系數(shù)來提高迭代的收斂速度。
顯示了系統(tǒng)的控制輸入狀態(tài)性能。閉環(huán)系統(tǒng)的輸入信號是有界的,證明筆者設計的控制算法具有良好的控制性能,達到了設計指標。
僅采用迭代學習控制算法對機械臂系統(tǒng)進行位置和速度跟蹤,給出了各關節(jié)的仿真結果如圖9~12所示,可以看出,控制效果較差,抖振現(xiàn)象較為明顯,且關節(jié)1的位置跟蹤和速度跟蹤始終存在誤差。
對比分析實驗結果可以看出,筆者設計算法的跟蹤控制效果較為滿意,控制系統(tǒng)具有良好的魯棒性,能夠有效抑制不同時變擾動,并提高了控制精度,實現(xiàn)了有效的軌跡跟蹤。
4 結束語
針對非線性不確定機械臂系統(tǒng)的輸出跟蹤問題,提出基于指數(shù)增益迭代學習的二階滑??刂品椒?。首先采用二階滑??刂葡饨绶侵貜托愿蓴_的影響,二階滑模方法可以使控制器產(chǎn)生連續(xù)的控制信號以削弱抖振效應,將連續(xù)二階滑模方法綜合到迭代學習控制系統(tǒng)的設計中對系統(tǒng)模型和有界外界擾動進行逼近,通過設置指數(shù)增益系數(shù)以提高迭代的收斂速度,在李亞普諾夫方法的基礎上,進一步研究和闡明了所提方法的收斂性和穩(wěn)定性。仿真實例表明,在不同時變擾動環(huán)境下,基于指數(shù)增益迭代學習的二階滑??刂凭哂辛己玫妮敵龈櫺阅堋?/p>
參 考 文 獻
[1] LIU J K,ER L J.Sliding Mode Controller Design for Position and Speed Control of Flight Simulator Servo System with Large Friction[J].Systems Engineering and Electronics(China),2003,14(3):59-62.
[2]? ?TRAN D T, TRUONG H V A, AHN K K.Adaptive nonsingular fast terminal sliding mode control of robotic manipulator based neural network approach[J].Precis Eng Manuf,2021,22(3):417-429.
[3]? ?PHAM D B,KIM J,LEE S G,et al.Double-loop control with hierarchical sliding mode and proportional integral loop for 2D ridable ballbot[J].Precis Eng Manuf,2019,20(9):1519-1532.
[4]? ?WANG H, PAN Y,LI S, et al. Robust sliding mode control for robots driven by compliant actuators[J].IEEE Trans Control Syst Technol,2019,27(3):1259-1266.
[5]? ?WU X R,HUANG Y Y.Adaptive fractional-order non-singular terminal sliding mode control based on fuzzy wavelet neural networks for omnidirectional mobilerobot manipulator[J].ISA Transactions,2022,121:258-267.
[6]? ?JD JESUS RUBIO.Sliding mode control of robotic arms with deadzone[J].IET Control Theory & Applications,2017,11(8):1214-1221.
[7]? ?CHEN M, YU J. Disturbance observer-based adaptive sliding mode control for near-space vehicles[J].Nonlinear Dynamics,2015,82(4):1671-1682.
[8]? ?INCREMONA G P,CUCUZZELLA M,F(xiàn)ERRARA A.Adaptive suboptimal second-order sliding mode control for microgrids[J].International Journal of Control,2016,89(9):1-37.
[9]? ?UTKIN V I,POZNYAK A S.Adaptive sliding mode control with application to super-twisting algorithm:Equivalent control method[J].Automatica,2013,49(1):39-47.
[10]? ?SHTESSEL Y,TALEB M,PLESTAN F.A novel adap-
tive-gain super-twisting sliding mode controller:Methodology and application[J].Automatica,2012,
48(5):759-769.
[11]? ?LEVANT A.Higher Order Sliding Modes and Their Application for Controlling Uncertain Processes[D].Moscow:Institute for System Studies of the USSR Academy of Science,1987.
[12]? ?LEVANT A.Sliding Order and Sliding Accuracy in Sliding Mode Control[J].International Journal of Control,1993,58(6):1247-1263.
[13]? ?KHAN M K,SPURGEON S K.Robust MIMO Water Level Control in Interconnected Twin-Tanks Using Second Order Sliding Mode Control[J].Control Engineering Practice,2006,14(4):375-386.
[14]? ?LU J, CAO Z,ZHANG R, et al.Nonlinear monotonically convergent iterative learning control for batch processes[J].IEEE Trans Ind Electron,2018,65(7):5826-5836.
[15]? ?BRISTOW D A, THARAYIL M, ALLEYNE A G.A survey of iterative learning control[J].IEEE Control Systems Magazine,IEEE,2006,26(3):96-114.
[16]? ?ARDAKANI M M G, KHONG S Z, BERNHARDSSON B.On the convergence of iterative learning control[J].Automatica,2017,78:266-273.
[17]? ?SHEN D,WANG Y.Survey on stochastic iterative lea-rning control[J].Process Control,2014,24(12):64-77.
[18]? ?WANG Y, GAO F, DOYLE F J. Survey on iterative learning control,repetitive control,and run-to-run co-ntrol[J].Process Control,2009,19(10):1589-1600.
[19]? ?LIU X,LI Y.The convergence of iterative learning control for some fractional system[J].Advances in Difference Equations,2017,132:4221-4360.
[20]? ?ONIZ Y,KAYNAK O.Control of a direct drive robot using fuzzy spiking neural networks with variable structure systems-based learning algorithm[J].Neurocomputing,2015,149:690-699.
[21]? ?XU J X,CAO W J.Learning variable structure control approaches for repeatable tracking control tasks[J].Automatica,2001,37(7):997-1006.
[22]? ?CHEN W, CHEN Y Q, YEH C P.Robust iterative learning control via continuous sliding?鄄mode technique with validation on an SRV02 rotary plant[J].Mechatronics,2012,22(5):588-593.
(收稿日期:2023-02-07,修回日期:2023-03-22)
The Second-order Sliding Mode Control Method for Manipulator
Based on Exponential Gain Iterative Learning
WANG Ce1, YANG Sheng2, ZHANG Lei1, ZUO Xin1, BAI Xiao-bo1
(1. College of Information Science and Engineering, China University of Petroleum (Beijing);
2. PetroChina Northwest Marketing Co.Wuhan Branch)
Abstract? ?Considering manipulator systems nonlinearity and uncertainty and its poor precision tracking control, a second-order sliding mode control algorithm based on exponential gain iterative learning was proposed. Firstly, having the sliding mode control used to enhance robustness of the system against external non-repetitive disturbances and the second-order sliding mode adopted to make the controller generate continuous control signals to weaken chattering in the traditional sliding mode control; secondly, having the exponential gain iterative learning control used to approximate the bounded repetitive disturbances caused by periodic disturbances such as friction and the exponential gain coefficient set to improve convergence speed of the iteration. This method can reduce tracking error of the system and ensure global asymptotic stability of the system as well as improve dynamic performance of the system; finally, having the sufficient conditions for convergence of the algorithm analyzed, and the theoretical numerical simulation of the proposed method carried out through taking a two-degree-of-freedom manipulator as an example. The simulation results show that, the control method proposed can ensure stable operation of the system, and the output tracking error converges to the neighborhood near zero to improve control accuracy of the system and maintain good robustness against interference.
Key words? ?two-order sliding mode control, iterative learning control with exponential gain, manipulator system,? trace tracking, chattering, control precision
作者簡介:王策(1997-),碩士研究生,從事控制工程的研究。
通訊作者:左信(1964-),教授,從事計算機模式識別、先進控制理論與應用和智能控制的研究,zuox@cup.edu.cn。
引用本文:王策,楊升,張磊,等.基于指數(shù)增益迭代學習的機械臂二階滑模控制方法[J].化工自動化及儀表,2023,50(5):644-651.