基于FPGA的非线性预测控制器设计

Implementation of Nonlinear Model Predictive Control Based on FPGA

对于非线性程度较高的复杂对象,非线性模型预测控制(Nonlinear Model Predictive Control,NMPC)是一种有效的控制策略。为了实现对这类对象的有效控制,设计了一种基于FPGA(Field Programmable Gate Array)的非线性预测控制器,该嵌入式控制器具有灵活性和高适应性等特点,能够应用于工业现场控制。为了满足工业控制的可行性和实时性要求,提出了一种序贯二次规划(SQP)算法的改进算法,在FPGA有限的计算资源下,保证每个采样间隔内都能得到NMPC优化问题的可行解。经仿真实验证明,采用非线性预测控制器在计算速度和精度上都能达到较好的性能。

Nonlinear model predictive control (NMPC) is an effective control strategy where application of nonlinear optimization methods is essential. In order to achieve effective control of such objects, this paper studies the implementation of nonlinear model predictive controller on FPGA, which, with high flexibility and adaptability, can be used in industrial control. In order to meet the feasibility requirements of industrial control, this paper propose an improved SQP algorithm, which could be applied on FPGA with limited computing resources, to get the feasible solution of NMPC optimization problem in each sampling interval. The simulation results show that the controller with algorithm can achieve good performance on both speed and accuracy.