多尺度卡尔曼滤波探测气体浓度

针对差分吸收激光雷达系统,研究了对原始测量信号和气体浓度模型采用多尺度分解 与卡尔曼滤波相结合的方法,实现了在强噪声的情况下对气体浓度进行探测。应用MonteCarlo仿真实验验证了该算法的有效性,仿真表明:该算法的性能优于仅在原始尺度上进行卡尔曼滤波的性能,也优于小波变换软阈值去噪的性能。

Multiscale Kalman Filtering Gas Concentration Detecting

As differential absorp tion lidar (D IAL) systems are affected by various noise factors, themeasured concen tration is corrup ted by noise, and can not be detected . ForD IAL systems , a multiscale Kalman filtering concentration detecting algorithm is introduced which uses wavelet transform as means of mapp ing data between different scales.Monte Carlo simulation shows that this algorithm is effective for detecting gas concentration in strong noise environ ment , and the performance of this algorithm is better not only than traditional Kalman filteringwhich processes at the original scale, but also than wavelet transform which has de-noise by soft-thresholding.