| Abstract [eng] |
The solution of lighting control problem is one of the most challenging tasks being encountered while developing the intelligent control systems. Primarily, the relevance of this problem was identified in the business sectors in order to reduce the indoor lighting costs. Nowadays, the other priorities have been identified, including the assisting control, health friendly environment, adaptation to changing human’s needs and individual habits. The major challenge while solving the lighting control task is to identify and implement significant requirements, because they determine not only the properties of the product, but its prevalence as well. The system, which focuses on lighting intensity standards, predefined control rules, usually does not pay attention to particular human needs and habits. On the contrary situations, when only human needs are estimated, the sufficient (health friendly) and cost-effective illumination is not always ensured. Finding the balance between human’s preferred lighting and recommended illumination may be beneficial solution. Besides, intelligent system has to be flexible and respond to changing environment, so human activity monitoring must be constantly taken into account. This document and developed software represent the proposed solution: intelligent lighting control system, which is based on the human’s lightning control habits and recommended illumination for the current activity. The interactive lighting control algorithm has been proposed, which provides lighting according user’s needs and recommended illumination. In order to create unnoticeable and convenient lighting control system, activity recognition is mapped to the image classification task, which allows to avoid of using various wearable sensors. The implementation of human activity recognition relies only on the combination of video monitoring devices and machine learning algorithms most suitable for image recognition tasks. |
