渐消递推滤波辅助光谱分析最佳自适应算法

FADING RECURSIVE FILTERING AIDED SPECTROSCOPY An Optimal Adaptive Algorithm

研究了Ｋａｌｍａｎ渐消递推最佳滤波辅助光谱辨析。基于Ｋａｌｍａｎ滤波达到最优时其新息序列互不相关的性质，提出了一种新的渐消滤波─—最佳自适应算法。通过在线自适应地调整遗忘因子从而使滤波（器）在存在模型误差或受到外扰时仍保持收敛性和最优性。用于重叠峰的辨识，能取得更好效果。

The fading Kalman recursive optimal filtering is studied for spectroscopicresolution,Based on the nature of Kalman filter that the residual sequenees arenot correlated when the optimal gain is obtained, a new fading filtering-optimaladaptive algorithm is proposed and utilized. Through the on-line and adaptiveadjustment of the fading or forgetting factor,the convergency and optimality ofKalman filtering are improved using measured outputs or estimated results, evenwhen there exist model errors and/or when the system is affected by unmeasurableexternal disturbances.The algorithm developed is applied to overlapped peakresolution with good results.