基于排列熵的长时间序列复杂程度分析

Complexity Analysis of Long Time Series Based on Permutation Entropy

针对光纤陀螺温度漂移过程中,长时间序列的复杂程度分析问题,提出了一种新的排列熵计算模型。排列熵算法能够有效放大时间序列的微弱变化,且计算简单、效率高,在时间序列分析方面具有很好的效果。但时间序列长度对排列熵的影响较大,同时,长时间序列的排列熵算法效率较低。为了有效计算长漂移序列的排列熵,引入包络曲线思想并计算包络均值,提出了一种新的均值排列熵算法。实验结果表明,均值排列熵算法不依赖于光纤陀螺漂移数据序列长度,提高了计算效率,是一种稳定性强的非线性特征参数,能够有效衡量时间序列复杂程度。

To analyse the complexity of the long time series during the temperature drift of the fiber optic gyroscope,a new permutation entropy calculation model is proposed.Permutation Entropy(PE)algorithm can effectively amplify slight changes of a time series data and has advantages of easy computing and efficiency,which has a good application in time series analysis.However,the length of the time series has a greater impact on the permutation entropy,and the efficiency of the permutation entropy algorithm for long drift series is low.In order to effectively calculate the permutation entropy of long time series,the idea of envelope curve is introduced and the envelope mean is calculated,and a mean permutation entropy algorithm is proposed.The experimental results indicate that the mean permutation entropy algorithm can be applied without the influence of the time series length of fiber optic gyro temperature drift,which improves the calculation efficient.It is a nonlinear characteristic parameter with strong stability and can effectively measure the complexity of time series.