Recursive estimation of linear systems
研究了有理传递函数标量线性系统的参数递归估计，通过仿真比较两种经典算法与Hannan & Rissanen三阶段算法，提出基于Givens变换的高效计算方案，并证明了三阶段过程估计量的中心极限定理。
The estimation of parameters for a scalar linear system with rational transfer function is considered, the estimates being updated each time a new data point comes to hand. Two well-known algorithms are compared by simulation with a method related to the three-stage algorithm of Hannan & Rissanen (1982); a computation scheme using Givens transformations is proposed which makes the three-stage recursive algorithm economical. A central limit theorem is proved for the estimators obtained by the three-stage process.