An Intelligent Smart Street Light System with Predictive model

• Published in: 2020 International Conference on System, Computation, Automation and Networking (ICSCAN) pp. 1 - 4
• Main Authors: Suresh, M., S, Anbarasi. M., V, PraveenKumar, A, Mohamed Hasvak
• Format: Conference Proceeding
• Language: English
• Published: IEEE 03-07-2020
• Subjects: Accelerometer sensor; Arduino; Computational modeling; Intelligent sensors; Internet of Things; IoT; LDR; Predictive models; Sensors; Software; Street light; Wireless fidelity
• DOI: 10.1109/ICSCAN49426.2020.9262435

This paper proposes the development of intelligent street lights that minimizes the wastage of power by applying the method of reliable smart mangement proposal that is essential for a wealthy lifestyle for upcoming generations living in modern platform. The propsoed model involves three main stages namely maintenence stage, automatic adaptive ON/OFF control stage and prediction of electricity consumption stage. In order to apply this technique, a Wi-Fi module, Light dependent sensor (LDR), accelerometer and ultrasonic sensors are employed. The accelerometer sensor is utilized for identifying the pole inclination and report the emergency condition to control room or user. Then, LDR sensor is employed to turn ON-OFF of street lights with respect to atmosphere intensity level. This switching could be performed using ESP8266 based on the LDR. Ultrasonic sensor helps to sense the presence of any vechicle or person in a specified range, then light glows with 100 % brightness else, it is reduced and glow with only with the intensity of 60%. Besides, a predictive model based on imporved bayesian neural netwok (IBNN) model is applied. The need of IBNN model is to report the utilization of the power for a period of specific duration. The propsoed work eliminates the power wastage during night time if there is no vechicle or trespassers passing. Moreover, by applying these three sensors, an intelligent automatic light is developed using IoT frame work that is effective.