Forecasting the State of Charging Batteries on Board the UAV on the Basis of Neuro-Fuzzy Network Using

The aim of the article is to develop methodological principles for predicting the condition of batteries on board unmanned aerial vehicles on the basis of a fuzzy network. A method for predicting the condition of batteries, adapted to the conditions and features of unmanned aerial vehicle use, has been developed. It consists in determining the load voltage, load current, solar current, battery temperature on board the unmanned aerial vehicle during its flight in real time and comparing the obtained data with the reference characteristics stored in the memory of the onboard unmanned aerial vehicle microcomputer. Based on the fuzzy logic in the support and decision-making system, the condition of the batteries is predicted to form a recommendation for the continuation or termination of the unmanned aerial vehicle flight task.

I. Introduction and Problem Statement

Today, it is difficult to underestimate the popularity and importance of using unmanned aerial vehicles (UAVs). However, during the flight of the UAV may occur abnormal situations (NA), in which the state of the technical complex, its components and funds involved, as well as the flight conditions, are not provided by the regular operation program. Occurrence of emergency situations can be connected with the following reasons: unreliability of elements of onboard systems of the UAV and their failure; error of the operational personnel of the mission control center (MCC) when operating the UAV; imperfection of methods and algorithms of UAV control with PMU; incompleteness and inaccuracy of information about the UAV as an object of management; limited ability to control the UAV, etc. One of the main factors leading to the emergency is the unpredictable discharge of the battery (DB). To ensure the trouble-free flight of UAVs, there is a need to develop intelligent systems that use methods of operational monitoring of the state of the AB in real time, which will assess the current state of the AB and predict its duration.