The goal of adaptive antennas is to maximise signal to interference ratio. This is reached by the use of adaptive algorithms. In recent time, their analog form is used, because it provides more simplicity and low price of the realisation in comparison with digital forms. These analog algorithms can be realised by the use of analog neural networks.

The submission deals with the analysis and solution of some problems of analog neural networks for the control of adaptive antennas. The substance of the solution consists in the control of the gain of closed loops with opamps in the analog circuit by the error signal, with the aim to avoid the dependence of convergence rate on adaptation parameters and input signals.