Abstract: The paper presents a model predictive controller based on the augmented non-minimal state-space model in stabilized platform according to the poor robustness of traditional model predictive controller and the serious complexity of fuzzy PID control system. Considering the actual situation a mathematical model of the stabilized platform generalized plant is established. Through the combination with receding horizon control principle and linear quadratic optimal control, an augmented non-minimal state-space model predictive controller is designed on the basis of the augment form of augmented non-minimal state-space model. An algorithm of state feedback gain matrix with Laguerre function is proposed, and simulation study is realized with MATLAB for steering drilling stabilized platform control system. The simulation results show that stabilized platform predictive control system can well satisfy the requirements of the drilling engineering for control precision and dynamic characteristics, and the system has strong robustness to the model perturbation and the disc valve friction disturbing torque with time variation.

Key words: stabilized platform, predictive control, non-minimal state-space, Laguerre function

摘要： 针对传统的模型预测控制器鲁棒性较差及模糊PID控制系统比较复杂的问题，提出了利用增广非最小状态空间模型与模型预测控制相结合的稳定平台预测控制。建立了稳定平台广义被控对象的数学模型，以增广形式的非最小状态空间模型为基础，结合滚动时域控制原则和线性二次型最优控制，通过对稳定平台离散模型的非最小状态空间形式进行增广变换，给出了基于Laguerre函数的状态反馈增益矩阵算法，设计了增广非最小状态模型下的预测控制器，实现了对导向钻井稳定平台控制系统的仿真研究。仿真结果表明稳定平台预测控制系统可以很好地满足钻井工程对控制精度和动态特性的要求，而且对有时变性的盘阀摩擦干扰力矩及模型参数摄动具有较强的鲁棒性。

关键词: 稳定平台, 预测控制, 非最小状态空间法, Laguerre函数