Neural signal processing systems with interdependent adaptation of components for noise immunity enhancement of radio communication system

Abstract

The article considers an data transmision system upgrade. It is based on one of the spectrum extention methods of transmitted signal that is the frequency-hopping spread spectrum. Analysis of the data transmission systems development shows that the main trend of changes at a level of data processing and transmission is the qualitative complication of physical organization for realization of new methods of adaptive signal processing. The use of a cellular neural network with hysteresis to perform state analysis of available frequency channels is proposed to increase the security of transmitted information. The results of neural network functioning allow predicting possible occupancy of each available radio channel and the probability of jamming at the moment of information transfer at a certain carrier frequency. This information allows adjusting the pseudo-random sequence generator operation generating the frequency partitioning sequence of available wireless communication channel. The performance in the form of adaptive neuristor lines that process signals transmitted between neurons is proposed to increase efficiency of the cellular neural network functioning communication between neural elements. The neural network system is able to process information not only in neurons with hysteresis, but also in the space between neural elements, which allows modifying the behavior of a neural network depending on the dynamics of signal propagation between its elements. Further the issue of variable parameters control of the designed complex neural structure by means of the modified algorithm of diffusive iterative search is considered. Application of similar optimization strategy leads to formation of metaheuristics which is capable to perform successfully tuning and training of a neural network complex. Check of network ability to predict frequency intervals with the minimal interference has shown that the cellular network is capable of performing reorganization of working frequency with account for features of background noise occurrence and jamming support in available frequency channels.