Performance Survey of Several Lossless Compression Algorithms for Telemetry Applications

Abstract

This paper presents a tutorial overview of several classes of known algorithms for lossless data compression and compares them based on their efficiency for compressing telemetry data. Performance of the various algorithms are evaluated, based on parameters such as compression ratio and processing time, using different test telemetry data files. The lossless algorithms considered are categorized into the following seven groups: suppression, substitution, bit level, relative, statistical, combination, and prediction, which includes both classical and neural predictors. Among the algorithms considered in this paper, it is found that prediction-based adaptive methods are ideal for real-time applications as they achieve good compression at relatively higher speeds. However, for off-line processing where processing time is not very critical, combination methods such as sixpack and Lempel-Ziv with Arithmetic Coding (LZARI) are better suited as they have the highest compression ratios. A twostage compression method involving a predictor and an encoder can be considered a trade-off between the former two schemes as this method achieves good compression ratios with reasonably low processing times.