Y.А. Bogatyrev - Ph.D. (Eng.), professor, radio receivers department, National Research University "MPEI". E-mail: bogatyrevya@mpei.ru
P.V. Zhiltsov - competitor for candidate of technical sciences degree, JSC «NAVGEOCOM». E-mail: pjghost@mail.ru

Relevancy. At present a great number of various wireless data transmission systems (DTS) are in use. Radio band is fully occupied with existing DTS and the problems in creating and using new DTS are still important. To the first place come the issues of increasing noise stability and efficiency for SPD which first of all are defined by radio signals properties. Modern data transfer systems uses different digital modulation schemes of a signal. Combining various methods of modulation it is possible to create a number of signals with different properties that are often opposite. Therefore the selection of the most suitable signals is complicated. There are also other factors and influences to be considered to ensure the maximum objectivity in proper signal selection. The objective of this work is to develop an efficient procedure for optimal signals selection that uses various quality performances and increases objectivity of a choice. By the optimal signal we mean the best signal that corresponds to the systems requirements. 15 different modulation schemes are analyzed in this work: PSK-4, DPSK-4, π/4-DPSK-4, QAM-4, QAM-16, FSK-4 1/T, FSK-4 1/2T, OFDM-4, CPM-4 Lin 0.5, CPM-4 Lin 0.25, CPM-4 Sin 0.5, CPM-4 Sin 0.25, MSK, GMSK 0.3 and GMSK 1000. To compare signals 7 quality performances (QP) which reflect various signal properties are used. Two solutions for optimal signal selection are offered: 1) In case of known values of the influencing factors, for each signal we should get the property estimation (for example, noise stability estimation). The signal with the greatest (or the smallest, depending on gradient set) value will be optimal. The estimation can be calculated either using known formulas or by computational modeling via computer simulation. An example of noise stability calculation via computer simulation is shown in this work. Several simultaneous influences are considered there, they are: AWGN and impulse noise, input filter gain-frequency characteristic of the receiver, output amplifier nonlinearity of the transmitter. MSK was determined as the optimal signal for these influences. Conclusions are drawn on change of a noise stability in case of different influences. 2) In case of unknown values of the influencing factors we need multi criteria selection procedure for optimal signals finding that can consider a set of QP and different requirements to DTS. In our work we apply the adaptive combined SπL-procedure which separates an initial set of QP {kl} (l = {1, M}) to r (1 ≤ r ≤ M) subsets with noncomparable QP inside. But the subsets are formed up by the linear order relation of their priorities. Such separating can be described with S-, π- and L-combinations as shown below: L (Ω/). As an example of introduced procedure using we want to select optimal signal for designed data transmission systems with next requirements: no special requirements. From problem situation it is impossible to get priorities between 7 QP. Thus we should add suprasystem metacriterias - imagined Cost and Efficiency of every QP. Then we used expert evaluation method to define every QP evaluation in these metacriterias and by this we determined criterias order (QP). The step-by-step solution of QP criteria order allowed us to define an optimum signal - MSK.

 

1. PodinovskijV.V. Mnogokriterial'ny'ezadachisuporyadochenny'mipovazhnostikriteriyami // Avtomatikaitelemexanika. 1976. № 11. S. 118-127.
2. Gutkin L.S. Optimizacziya radioe'lektronny'x ustrojstv po sovokupnosti pokazatelej kachestva. M.: Sov. radio. 1975. 368 s.
3. Kandy'rin Ju.V. Metody' i modeli mnogokriterial'nogo vy'bora variantov v SAPR: Uchebnoe posobie dlya vuzov. M.: Izd-vo ME'I. 2004. 172 s.
4. Molodczov O.A., Fedorov V.V. Ustojchivost' princzipov op­timizaczii: Sovremennoe sostoyanie teorii issledovaniya operaczij. M.: Nauka. 1979. S. 236-263.