抄録
We have developed a compression parameter tuning algorithm for the piecewise linear compression algorithms, such as the Boxcar Algorithm, Swinging Door Algorithm and a newly developed compression algorithm. The developed tuning algorithm requires only the targeted compression factor, and it tunes each compression parameter according to the statistical analysis of the first order or the second order differential of the source signal. The test result obtained using simulated signals show better ability to compress the signals with Brownian motion, Gaussian noise and 1/f noise by the developed compression algorithm and the Swinging Door Algorithm compared to the Boxcar Algorithm. Against the real signals, the developed algorithm achieved better compression ability than others especially toward the smoothened signals.
