线性系统集员估计问题研究

Research on Set-Membership Estimation for Linear System

研究了线性系统下的Norton和基于OBE两种集员估计算法。Norton算法是通过最小体积或最小迹来优化时间更新阶段和测量更新阶段,但其计算量大、效率低。针对这一不足,OBE算法采用最小半径定界椭球来进行测量阶段的更新,从而简化了算法,减少了计算量。最后通过与传统Kalman滤波算法与Norton集员估计算法相比,验证了基于OBE集员估计算法的有效性。

This paper studies two set-membership estimation algorithms in the linear system： Norton set-mem- bership estimation algorithm and OBE set-membership estimation algorithm. Norton algorithm optimizes time updating stage and measurement updating stage according to the minimum volume or minimum trace, but with huge amount of computation and low efficiency. OBE algorithm optimizes measurement updating stage according to minimum radius bounding ellipsoid, so as to simplify the algorithm and reduce the amount of calculation. Finally, OBE algorithm is compared with the traditional kalman filtering algorithm and Norton set-membership estimation algorithm. The simu-lation results prove the effectiveness of OBE algorithm.