A Lossless Compression Method of Time-Series Data Based on Increasing Average of Neighboring Signals

Abstract

Golomb-Rice coding is a well-known compression algorithm for sensor data. When time-series data changes drastically with the large amplitudes such as a pulse signal, the code length based on Golomb-Rice coding becomes very long. In order to shorten the code length, amplitude of signal is decreased by calculating differential signal between a raw signal with a similar signal. In this paper, we develop a lossless compression method for time-series data such as sensor data. In traditional methods, finding the past-signal from which a differential signal with low amplitude can be generated is the main topic. However, if there are no past-signals to reduce sufficiently the amplitude of differential signal, the data compression procedure takes only low effects. In our approach, a signal which decreases energy of a pulse signal or increases energy of the neighboring signal of a pulse signal is adopted to generate differential signals. In order to select an effective signal, we propose a method for detecting reference signals based on cumulative distribution features of time-series data. As results of experiments, we confirm that our proposed method can generate codes whose length is shortened. The code length was decreased to 97% on average and up to 81% in comparison with the traditional method.