Hardware Interfacing Library for Angle-of-Arrival Estimation in Applied Artificial Intelligence Systems

This article discusses a set of issues related to the implementation of an auxiliary library for interfacing hardware platforms for radio signal acquisition and versatile computer-based signal processing systems aimed at using intellectual approaches such as machine learning and artificial neural networks (ANN). The architecture of the system is considered, including various layers of abstraction for collecting, storing, multistep processing and visualizing information, taking into account the essential features of working in two modes - research and production. A detailed description of the main modules of the system, the principles of their interaction and the position of the intelligent data processing unit in the processing pipeline are provided. The feasibility of the proposed approach was assessed using the example of solving the problem of angle-of-arrival estimation using a pre-trained ANN of modified multilayer perceptron architecture. The paper provides a detailed description of the ANN architecture, the formation of its input data vector, and presents the results of the numerical simulation.