The Lower Boundary Capacity Estimation of Frequency Selective Channel Utilizing PAM-n-signal with Independent and Identically Distributed Values Channel Symbols Part I. Standard partial signals

This paper is devoted to the study of Faster than Nyquist regime in frequency selective channel with PAM-n-signals utilizing the resolution time theory, which is used to estimate the duration of a channel symbol for given value of BER. To achieve the goal of paper, the original math model of the information channel presented in previous authors papers was modified, which is more convenient from the point of view of obtaining results suitable for engineering analysis. Thanks to the modified model, the following equalities are obtained: the equality that allows in terms of the lower bound of capacity estimation (according to Sh. Shamai); the minimum value of the signal-to-noise ratio at which the required value of BER is achieved; the estimation rule of optimal time sampler shift. Using the obtained relations, numerical estimations of the indicated parameters were obtained for the most typical types of frequency selective channels, which is simulated by Butterworth filter type. New properties have been revealed that indicate the possibility of achieving a high specific capacity of 4.5 bits/Hz*s with SNR = 25 dB and BER = 10
⁻¹²