提出利用功率谱密度分析、自相关函数分析、Allan方差分析等方法建立可靠的钟差函数模型和随机模型的思路,并对比分析了现有的两种两参数钟差状态模型——随机游走型状态模型和高斯-马尔可夫型状态模型。计算结果表明,在GPS/INS紧组合导航应用中,应使用高斯-马尔可夫型钟差状态模型,并通过自相关分析等方法估计可靠的相关时间,方能准确地描述钟差状态变化。
A method to constructe a reliable functional model and a stochastic model for clock error based on the theory of refining filter model by time-frequency analysis, power spectral density, autocorrelation and Allan variance was proposed. Comparison of the results calculateed with random walk type and with Gauss-Markov type shows that only by using Gauss-Markov type clock model together with the autocorrelation analysis of correlation time, the variation disciplinarian of clock can be represented properly.