具有不确定测量输出系统的滚动时域估计

Moving horizon estimation for stochastic systems with missing measurements

针对具有不确定测量信息的系统讨论了一种滚动时域估计(moving horizon estima-tion,MHE)方法。首先,在已知的错误信号输出概率的基础上,融合预测控制的滚动优化原理,在每个采样时刻通过极小化优化问题的性能指标估计出系统的初始状态和作用在系统上的扰动,接着再由系统的动态方程计算状态的估计值。仿真结果表明:与卡尔曼滤波方法相比,MHE方法能处理系统约束,具有比卡尔曼滤波更好的估计性能。

Considering estimation （MHE） a linear uncertain system with strategy was proposed. Based on missing measurements a moving horizon the probability of occurrence of missing measurements, the initial state and disturbance were estimated by minimizing performance object at sample time, and then the estimation value was calculated by dynamic equation of the system. The simulation results show that MHE can solve the constrained linear system and has more effective estimation performance compared with Kalman filter strategy.