Development of a control automation platform based on artificial intelligence and machine vision using neural networks

Abstract

For modern industrial production and business processes, automation plays a key role in increasing efficiency, reducing costs, and improving product quality. The use of artificial intelligence (AI) and machine vision opens up new possibilities for control automation in various fields. This paper is devoted to the development of a control automation platform based on neural network algorithms integrated with machine vision methods in order to enhance the functionality and versatility of automated systems. The developed platform is based on a deep learning architecture that includes convolutional neural networks (CNNs) for processing visual data. The main focus is on integrating computer vision techniques with control modules that use recurrent neural networks (RNNs) to analyze time sequences of data. Simulation and experimental settings were used for testing to evaluate the accuracy, speed and adaptability of the system being developed. Neural networks were trained on large amounts of data using pre-prepared sets of images and real data from production facilities. The developed platform has shown high accuracy in recognizing objects, tracking movement, and making decisions in the face of changing environmental parameters. Compared to traditional control algorithms, the neural network-based system has demonstrated a 30% increase in performance, a 20% reduction in processing time, and increased operational stability through self-learning. The results obtained confirm the effectiveness of the proposed platform and its practical applicability. However, areas for improvement have been identified related to optimizing computing resources and eliminating possible errors in data processing. The developed AI and machine vision-based platform offers a promising management automation solution that facilitates the integration of technologies into industrial and commercial processes. Further steps include improving the models and expanding their functionality.