Iнформаційно-управляючі системи біотехнічними об’єктами з інтелектуальними підсистемами

Abstract

Проаналізовано існуючі архітектури та системи контролю біотехнологічних виробництв, запропонувано власну архітектуру нейронної мережі для підтримки модулів прийняття рішень. Для усунення шумів інформаційного каналу розроблено фільтр на основі перетворення Гільберта-Хуанга.

The existing architecture and biotechnology control facilities systems were analyzed in the article. The neural network architecture is offered to support decision-making units. To eliminate information channel’s noise components the Hilbert-Huang transformation filter was designed. Improvement of biotechnological objects systems’ architecture management was reached by including neural network blocks into intelligent subsystem decision-making processes due to the content of biological objects. This will expand the list of effectively processed natural disturbances and ensure the sale profit increase of the products produced with minimizing energy costs for maintenance. The combination of recognition systems and Bayes networks lead to the sensitivity rising before image changing. Due to the fact that information and control systems architecture of biotechnical facilities was added to neural network forecasting unit, when dealing with stationary (quasi-stationary) time series, we managed to achieve sufficient depth forecast. However, this approach did not work in non-stationary time series. This led to the before filter input signal usage. Overall, the effectiveness of such control systems depends on the functional characteristics of two components: biological and technical.