Matrix model parametric identification of antenna tuning units

Abstract

HF radiocommunication modern development trends provide for improved handling and increased speed of antenna tuning units (ATU) functioning. The need to control the tuning process using a microcontroller based on the measurement results of antenna input impedance in the “on-line” mode requires an increase in the accuracy of modeling the transformative properties of these devices and subsequent calculations for the development of control commands by discrete tuning bodies. The more reliable the data in the memory of the microcontroller will characterize the real properties of the device's tuning bodies, the better their regulation will be. Based on the results of a full-scale experiment, the issues of parametric identification of the previously proposed conceptual matrix model of discreet power circuit are considered. Due to the complexity of technological modeling of circuits transformative properties, the task is to increase the accuracy of its numerical results for each device. With the found set of solutions for the model with the selected structure and a significant number of unknown parameters, its identification by the objective function extremum is not possible. Accordingly, the analysis of a priori identifiability (theoretical, with unambiguous definition of parameters) is carried out on the basis of a qualitative study of the structure of the model and the scheme of the “ideal” experiment (with determinism of all unmistakably measured values). The identifiability of the proposed power circuit model has been confirmed. The variants of the experiment, the algorithm for processing its results, as well as ways to obtain their guaranteed quality with the help of a vector circuits analyzer for a certain time at a given cost are described. The criterion for assessing the conformity of the model to a real power circuit is a measure of the number of its errors after measuring complex coefficients on a set of frequencies. Processing the results using the least squares method with interpolation for the entire frequency band reduces the errors and complexity of the experiment. Modeling of a 9-bit matching inductive body confirmed the correctness and effectiveness of the matrix description of the transforming properties of the wide band ATU and other solutions presented, as well as the possibility of full automation of this experiment.