| 摘要 |
PROBLEM TO BE SOLVED: To conduct signal decoding with small error with a simple means and to make the signal processing efficient by using a predicted value decided based on a real sampling value in place of part of fractions other than the real sampling value and decoding the data so as to process the signal with the data. SOLUTION: A continuous signal f(t) being an object of processing is sampled at a period T and a sampling series is generated. A sampling predict means predicts a sampling at a plurality of m set sampling points other than the real sample value in a small number n set of sampling points obtained by sampling the continuous signal. Furthermore, a sampling value at a non-sampling position (t=t') is decoded by calculating convolution integration between discrete signals and sampled functions at (n+m) points consisting of the real sampled value and the predicted sampling value. |
