摘要
针对NARX模型中一步时滞控制算法不完全适用于多重时滞的情形,提出一种在工作点处用线性的动态切平面逼近一般的非线性系统的方法,结合一种准则函数和预测器算法,提出一种自适应预测控制算法,给出了该模型下具有遗忘因子的递推最小二乘非线性参数估计算法。仿真结果表明:该控制算法品质好,参数估计算法几乎无偏,验证了自适应预测控制算法和递推参数估计算法的有效性。
Directed at addressing the poor application of many one step time delay control algorithms in the NARX model to multiple time delay situation, the paper features an approach which consists of the use of linear dynamic tangential planar to make an approximation generic nonlinear system on operating particle, the combination of a criterion function and the predictor algorithm, the development of an adaptive prediction control algorithm, and the introduction of recursive least squares nonlinear parameter estimation algorithm which owns forgetting factor under this model. Simulation results show the control algorithm is of better quality and the parameter estimation is almost unbiased. The adaptive predictive control algorithm and recursive parameter estimation algorithm prove to be valid.
出处
《黑龙江科技学院学报》
CAS
2008年第2期137-139,143,共4页
Journal of Heilongjiang Institute of Science and Technology
