Neural Network-Based Intelligent Lighting System for Indoor City Farms: Architecture, Implementation, and Energy Efficiency Analysis

Autores

  • Natalia Axak Kharkiv National University of Radio Electronics
  • Yurii Shelikhov Kharkiv National University of Radioelectronics
  • Maksym Kushnaryov Kharkiv National University of Radio Electronics

Palavras-chave:

intelligent lighting, IoT, artificial neural networks, energy efficiency, cloud computing, urban farming, machine learning

Resumo

Urban agriculture is gaining importance as cities seek sustainable solutions to food production challenges in limited spaces. Artificial lighting is critical in indoor farming, yet conventional systems based on fixed schedules or manual control are energy-inefficient and fail to respond to real-time plant needs. This paper presents an intelligent lighting system for indoor city farms that leverages a cloud-hosted artificial neural network (ANN) and real-time sensor data to optimize illumination dynamically.

The proposed system integrates temperature, humidity, and light sensors with low-cost microcontrollers and a cloud-based ANN model trained on 1,500 empirical and 3,500 synthetically generated data samples. The ANN predicts optimal light intensity (lux) based on current environmental conditions and plant growth stage.

Experimental deployment over three months demonstrated up to 29.2% energy savings and a 24% yield increase compared to manual and timer-based systems. Unlike prior work, this system provides a real-time, closed-loop architecture combining edge-level sensing and cloud intelligence. The approach supports scalable, modular deployment for both home and commercial vertical farms and contributes to advancing sustainable, data-driven agriculture.

Downloads

Não há dados estatísticos.

Publicado

2025-12-05

Como Citar

Axak, N. ., Shelikhov, Y. ., & Kushnaryov, M. . (2025). Neural Network-Based Intelligent Lighting System for Indoor City Farms: Architecture, Implementation, and Energy Efficiency Analysis. Anais Do Congresso Internacional De Conhecimento E Inovação – Ciki, 1(1). Recuperado de https://proceeding.ciki.ufsc.br/index.php/ciki/article/view/1740