多变量CARMAX模型的在线修正参数预测滤波PID控制  

Prognosis-filtering PID control with on-line modifying parameter for multi-variables CARMAX model

作　　者：侯小秋[1] 

机构地区：[1]黑龙江科技大学电气与控制工程学院,哈尔滨150022

出　　处：《黑龙江科技大学学报》2016年第1期68-74,共7页Journal of Heilongjiang University of Science And Technology

摘　　要：针对多变量CARMAX模型,采用系统的输出预测值代替系统的当前输出值,提出具有预测控制性能的增量型预测滤波解耦PID控制算法。采用可克服病态的多变量系统的遗忘因子递推最小二乘算法对被控对象进行参数估计,利用自校正预报显式算法对系统输出进行预测,根据可克服算法病态的直接极小化指标函数自适应控制算法和Robbins-Monro算法,给出了具有在线修正PID控制参数和加快PID控制参数收敛性能的多变量CARMAX模型的自适应预测滤波解耦PID控制算法。仿真结果表明:改进的PID控制算法具有预测控制性能和在线修正参数性能,系统具有较好的控制品质。This paper proposes a novel algorithm for prognosis-filtering-decoupling PID control with on-line modifying parameter——an algorithm developed by using the forecast output values of the system instead of the current output values of the system in response to multi-variable CARMAX model. The algorithm study involves developing a multi-variable forgetting factor recursive least squares algorithm with solving ill-controlled; estimating the parameter of the controlled model by employing the algorithm; and developing an adaptive prognosis-filtering-decoupling PID control algorithm with the characterizations of on-line modifying parameter and speeding the convergence of PID control parameter for CARMAX Model with multivariable by employing the self-tuning prediction explicit algorithm to predict the system output,and drawing on the adaptive control algorithm with solving ill-controlled for direct minimization index function and the Robbins-Monro algorithm. The results indicate that the improved algorithm capable of prediction control and on-line parameters correction offers better control quality.

关 键 词：自适应控制 预测控制 PID控制 参数估计 自校正预报 多变量系统 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]