三、状态滤波 前面我们已经用动态规划的方法获得了最佳控制函数: U_k~*=-ΦX_k其中Φ是状态方程中的转移矩阵,是离线计算所得的最佳反馈系数矩阵。在线控制时,依据每一时刻的量测所决定的状态X_k计算出每一时刻的控制量U_k~*,实现其最佳控制。但是由于系统存在着随机干扰,量测也存在随机误差,因此不能把量测值直接作为反馈的依据,而必须对状态进行滤波估计,把滤去了干扰的状态量作为反馈的依 State Filtering In the previous section, we have obtained the optimal control function by the method of dynamic programming: U_k ~ * = - ΦX_k where Φ is the transition matrix in the state equation and is the matrix of the best feedback coefficient calculated offline. Online control, according to the measurement of each moment of the state X_k calculated at each moment of control U_k ~ *, to achieve its optimal control. However, due to random disturbance in the system, there is random error in the measurement. Therefore, the measured value can not be directly used as the basis of the feedback, but the state must be filtered and estimated, and the state quantity with the interference removed as the feedback