AUTOMATIC SYSTEM OF SPEED CONTROL OF UNDERWATER VEHICLE BASED ON ARTIFICIAL NEURAL NETWORKS WITH REFERENCE MODEL

摘要

The invention relates to shipbuilding and can be applied at underwater vehicles which as an engine to move along an axis of coordinates use a propeller with electric drive, and which should have high performance and accuracy of speed control. An automatic system of speed control of an underwater vehicle based on artificial neural networks with a reference model comprises a setter of input signal, a speed sensor, series connected power converter, a motor with a propeller and the underwater vehicle as a control object, and is characterized in that it additionally comprises a first unit of output signal delay associated with the speed sensor, a second unit of output signal delay associated with the first delay unit, and a computer, connected by the inputs to the speed sensor, the first and second units delaying the output signal and the input signal setter, and by the output – to the power converter and equipped with a program based on the artificial neural networks, which is previously trained to reproduce together with the object reference model dynamics. The technical result of the present invention is to improve the accuracy of reproducing dynamics of the reference model, because all non-linearity of the object are not taken into account by means of mathematical modeling, but on the base of data obtained experimental directly from the object, to simplify the definition of the parameters of the underwater vehicle.