IoT-Based Road Vehicle Counter Using Ultrasound Sensor and Cross-Correlation Algorithm

Abstract

Traffic systems at this time have shown evolution as technology develops. Supported by a transportation system that is certainly sophisticated. The two systems are part of smart city that are applied in big cities. Basically, all communicate with each other to create an integrated smart city. However, the communication must be in real time domain so that all smart city components are connected. In this research case a vehicle counting system in real-time that can calculate vehicles passing on a road segment is designed. Applications used are ultrasound sensors, microcontrollers, and an Internet of Things Platform that are interconnected to monitor road conditions. Normalized Cross-Correlation algorithm is used to detect passing vehicles. The concept that Normalized Cross-Correlation algorithm is an algorithm to determine the similarity in two frequency signals is used to detect ultrasound frequencies created by cars passing by the sensor. The system will detect by comparing input data from ultrasound sensors by making sample data first then the sample data is compared with the data after the sample data. After that the correlation value will come out which has been normalized on a scale of 0–1.0. From applying normalized cross-correlation method the threshold for the calculation of the vehicle is determined, which is ¡0.70. This threshold is determined as the optimum value after various tests. After testing the method in real environment the error rate of the method in counting passing vehicles is 10.1%.