Design, Simulation, and Analysis of a Solar-Powered Street Lighting Control System for Power Consumption Prediction

This paper is devoted to designing, modeling, and analyzing a solar-powered street lighting system using artificial intelligence technologies to predict energy consumption. The system involves the integration of solar panels, batteries, and sensors to efficiently control the brightness of LED lamps depending on the illumination level and the presence of motion in the lighting area. The use of artificial neural networks (LSTM, GRU, Random Forest) provided accurate prediction of solar energy generation and the need to connect to the city power grid during periods of low solar activity. The developed model was implemented in MATLAB Simulink, considering actual weather data. The results showed that the system can operate autonomously for up to three days under unfavorable weather conditions, reducing power by 30%. The need for connection to the external network is no more than 10% in summer and up to 30% in winter. The prototype will be implemented in Kazakhstan to validate the model and further optimize its operation.