A Model for Non-Stationary Time Series and its Applications in Filtering and Anomaly Detection

Time series measurements from sensing units (e.g., UWB ranging circuits) always suffer uncertainties like noises, outliers, dropouts, and/or nonspecific anomalies. In order to extract the true information with high precision original corrupted measurements, signal-model-based signal pre-processing embedded in circuits are usually employed. However, for a general observe, its model cannot be obtained so that processing methods not applicable. this article, time-variant local autocorrelated polynomial (TVLAP) state space is proposed dynamics of non-stationary stochastic process (i.e., or time series), through which model-based could utilized denoise, correct outliers/dropouts, identify anomalies contained measurements. Besides, presented method can also used change point detection series.