Comparative Study of CNN–LSTM and ConvLSTM Models for Short-Term Power Forecasting in Smart Street Lighting Systems

Engineering notes

Engineering notes will be added by the Power for Solar editorial team.

Chinese explanation / 中文解读

中文解读待补充：本站会优先为光伏效率、钙钛矿太阳能电池、储能技术、太阳能热利用、BIPV、并网技术等高价值论文补充中文说明。

Original abstract

Smart street lighting systems (SSLs) have emerged as an effective solution for reducing energy consumption and promoting environmental sustainability by providing adaptive lighting that responds to user demand without compromising safety or security, thereby improving overall efficiency. The application of computational techniques, including machine learning and deep learning, is essential for enhancing smart street lighting systems. These methods enable the prediction of power consumption, allowing dynamic adjustment of lighting levels based on real-time demand. As a result, integrating advanced algorithms improves energy efficiency, reduces operational costs, and contributes to sustainable urban lighting.This study investigates the application of deep learning models, namely Convolutional Neural Network with Long Short-Term Memory (CNN–LSTM) and Convolutional LSTM (ConvLSTM), for predicting power consumption in a smart lighting system. A univariate time-series dataset of power consumption, collected over a seven-day period, was employed. The methodology utilized a one-step-ahead prediction strategy based on actual values. Experimental results showed that both models achieved satisfactory predictive performance.

Links and sources

• PDF from original source

Comments