基于卡尔曼滤波理论的气体浓度反演算法

Gas Concentration Retrieval Algorithm Based on Kalman Filtering Theory

广泛应用于气体探测的差分吸收光谱技术(DOAS)利用气体分子的窄带吸收特性结合最小二乘算法来推演气体浓度。但是,最小二乘法在外界环境因素干扰的情况下,往往产生较大的误差。本文引入了基于状态空间理论的气体浓度定量分析算法。通过把浓度变化视为状态方程,把光强吸收变化看作测量方程,从而组成状态空间方程,然后将卡尔曼滤波应用到气体状态空间中实现浓度反演。对于噪声统计信息未知的情况,通过自适应滤波算法,在滤波过程中利用已有的历史信息对噪声实现估计,从而使得整个系统在信噪比较低的情况下也能取得较好的反演精度。最后通过实验对最小二乘算法和卡尔曼滤波算法进行对比,证明卡尔曼滤波算法更具优越性。

The Differential Optical Absorption Spectroscopy （DOAS） technique uses the narrow molecular absorption bands and absorption strength to retrieve concentrations of the gases based on least square methods. We propose an algorithm for gas concentration recovery for the real world case where the DOAS system model is subject to uncertainties. The method is based on the formulation of concentration reconstruction as a regularization problem, where the dynamic change of gas concentration is treated as a state equation and the measurements of DOAS is used to form the measurement equation. The estimation is achieved with the aid of an adaptive Kalman filter framework. It is derived and extended from the Kalman filtering principles and is particularly powerful for real-world situations where the levels of the disturbances are unknown. The performance of our work is evaluated by using the experimental data.