Симуляційне дослідження енергоефективної системи керування світлодіодною матрицею з елементами граничного штучного інтелекту

Abstract

Simulation study of energy-efficient LED matrix control system with edge artificial intelligence elements. The paper presents an energy-efficient control system for LED matrix using edge artificial intelligence (TinyML). The system dynamically adjusts brightness, refresh rate and power modes based on recognized image context, significantly reducing energy consumption. All research was performed in simulation environment with Proteus, Edge Impulse and TensorFlow Lite Micro tools. Results demonstrate 94–96 % recognition accuracy, inference time under 80 ms, and power savings of 42–58 % (up to 70 % in static scenarios) compared to traditional fixed-mode control. The proposed approach is promising for autonomous IoT devices and wearable gadgets