精密单点定位的卡尔曼滤波新息与残差随机性对比

为了验证卡尔曼滤波新息和残差描述观测噪声的能力，根据精密单点定位载波相位和伪距观测值的先验方差，利用蒙特卡罗模拟等方差的随机噪声作为分析基准，研究描述信号随机性的平稳性、正态性、自相关性和功率谱指标。实验结果表明，新息只能正确描述伪距观测值噪声的随机性，残差能够同时描述载波相位和伪距观测值噪声的随机性。同时，两者描述伪距观测值噪声的能力相近。因此，不同精度观测值的精密单点定位应该以新息表示的残差为基础进行卡尔曼滤波模型的改进。

Contrast of Randomness in Innovation and Residuals of Kalman Filter of Precise Point Positioning

To verify the ability of the noise measurement described by Kalman filter innovation and residuals, four indexes for describing the randomness of signals, including stability, normality, auto-correlation and power spectrum were studied according to the priori variance of carrier phase and pseudorange of precise point positioning and utilizing the random noise with the same variance generated by Monte Carlo method as an analyzing basis. The experiment demonstrated that residuals could describe the randomness of the observation noise in both carrier phase and pseudorange, while innovation could only describe the randomness of pseudorange noise. Meanwhile, the abilities of carrier phase and pseudorange in describing the pseudorange noise are the same. Therefore, residuals expressed by the innovation are better for improving Kalman filter model in precise point positioning of measurements with different precisions.