【摘要】：本文針對一類線性離散時不變系統(tǒng),利用共軛方向優(yōu)化方法設(shè)計了一種迭代學(xué)習(xí)控制算法.首先,基于采樣數(shù)據(jù)構(gòu)建超向量,將原二維動態(tài)系統(tǒng)轉(zhuǎn)化為迭代域中的一維系統(tǒng).其次,在這種形式下,利用當(dāng)前的跟蹤誤向量減去其在以前搜索方向上的投影,構(gòu)建新的搜索方向,以補(bǔ)償當(dāng)前的控制信號,進(jìn)而構(gòu)建下一次迭代的控制信號.再次,結(jié)合共軛方向的性質(zhì),利用數(shù)學(xué)歸納法分析了算法的單調(diào)收斂性和二次終止性.最后,數(shù)值仿真驗證了理論分析的正確性和有效性;同時,與已發(fā)表的比例型和范數(shù)最優(yōu)迭代學(xué)習(xí)控制方法進(jìn)行比較,得出了本算法的優(yōu)越性.
[Abstract]:In this paper, an iterative learning control algorithm is designed for a class of linear discrete time-invariant systems by using the conjugate direction optimization method. Firstly, the supervector is constructed based on the sampled data, and the original two-dimensional dynamic system is transformed into one-dimensional system in iterative domain. Secondly, in this form, the current tracking error vector is used to subtract its projection in the previous search direction, and a new search direction is constructed to compensate the current control signal, and then the control signal of the next iteration is constructed. Thirdly, combined with the properties of conjugated direction, the monotone convergence and quadratic termination of the algorithm are analyzed by mathematical induction method. Finally, the correctness and effectiveness of the theoretical analysis are verified by numerical simulation, and the advantages of the algorithm are obtained by comparing with the published proportional type and norm optimal iterative learning control methods.
【作者單位】： 西安工程大學(xué)理學(xué)院;西安交通大學(xué)數(shù)學(xué)與統(tǒng)計學(xué)院;
【基金】：The Doctoral Foundation of Xi’an Polytechnic University(BS1617)
【分類號】：O231

【相似文獻(xiàn)】

相關(guān)期刊論文 前10條

1 李新忠,簡林柯,何擔(dān)，

本文編號：2476070