面向鲁棒控制的l_1中心估计递推算法

A Robust Control Oriented l_1 Central Estimation Recusive Algorithm

讨论未知但有界误差假设下的ｌ１中心估计问题，提出了中心估计递推算法的理论基础，并据此设计了ｌ１中心估计的递推算法。理论分析和数据仿真表明，与原有算法相比，递推算法可以显著减少计算量，提高估计速度，因而可应用于实时在线辨识。所给算法和结果是面向鲁棒控制的。

The problem in l1 central estimation with uncertainty but bounded measurement error is discussed in the paper. The recusive algorithm is proposed together with its basic theorem. The theoretical analysis and simulation results show the amount of calcutation can be reduced notably and the estimation process can be accelerated. Therefore, this kind of recusive algorithm can be used in on-line identification. The algorithm and results proposed here are robust control oriented.