A.A. Gladkikh – Ph. D. (Eng.), Associate Professor, Professor, Department «Telecommunications», Ulyanovsk State Technical University
N.A. Pchelin – Post-graduate Student, Department «Telecommunications», Ulyanovsk State Technical University
On the basis of statistical modeling the paper shows the ability to use the patterns of changes in estimates of binary symbols soft de-cision as the executive instrument for the switching of the adaptive modes of data exchange system. The essential element of any adaptive system is a subsystem of estimation of the external environment parameters. When transmitting data over the continuous channels with non-stationary characteristics the estimation of the noise-level, as a rule, is carried out only on the data link layer by registration of the errors during the decoding of used error-correcting code combinations. Obviously, this approach introduces a noti-ceable time delay in the adequate response on the actual changes of environmental parameters observed as a change of characteristics of the communication channel. In order to improve the energy efficiency of the modern data exchange systems, the soft methods of redundant codes processing are being actively implemented. Herewith, in the vast majority of sources it is discussed the application of soft binary symbol decision as the improvement the correcting capabilities of redundant codes. For the best of our knowledge the application of soft binary symbol decision as the improvement the functioning algorithms of adaptive data exchange systems was not considered so far.
As the sensory characteristics this paper investigates the change in the mathematical expectation of soft symbol decision estimates and their variance. It is shown that the mathematical expectation is more stable while change of the signal-to-noise and therefore can be recommended as a criterion for switching from one of the specified modes of adaptation to another. The series of statistical model tests prove that the tuple length of the scoring segment is not of a great importance to determine the average values of soft symbol decisions. It gives the opportunity to warn about the switch of system operation mode by using the short data tuples, which has a positive effect on the system’s dynamic characteristics. The change of the variance has a large range of its values with the fixed signal-to-noise. For this reason, this statistical parameter cannot be a criterion for switching adaptive modes. The paper also shows the technology of adaptation mode switching by the parameters of the redundant codes.