基于状态空间模型广义预测控制的并行算法

Parallel Algorithms for the Generalized Predictive Control Based on State-Space Model

中文关键词  系统辨识  广义预测控制  并行算法  systolic阵列

英文关键词  parameter estimation  generalized predictive control  systolic array  parallel algorithm.

中文摘要

      本文首先基于脉动（Systolic）阵列结构，提出了一种实时参数辨识的并行算法，然后推导出基于状态空间模型广义预测控制（GPC）的两种新算法，这两种算法都可以通过阵列结构并行实现。

英文摘要

      In this paper, an efficient algorithm and Systolic architecture are proposed for parameter estimation. Two algorithms and Systolic architectures for the generalized predictive control （GPC）based state-space model are developed. A new recursive solution of GPC is developed, in which the control input can be obtained recursively from discrete Riccati equation. The recursive algorithm can be implemented by a small size of array. Another algorithm is mapped to a set of two dimensions arrays for computing predictive control and output. The analysis shows that the methods have high degree of parallelism.