Article Preview
Top1. Introduction
In recent years, the urban population growth is increased tremendously. As per the recent statistics of the World Health Organization (WHO), From 2015 through 2020, the global urban population will expand by 1.86 percent annually. This growth is expected to be 1.63% between 2020 and 2025 and 1,44% between 2025 and 2030. In urban areas, a considerable proportion of cars are owned by a single home, and at least two cars are owned by a single home. Private cars are becoming increasingly popular with urban traffic. This means that transport in metropolitan areas all over the world is becoming one of the biggest concerns (Shafiq et al. 2020). The large majority of individuals trafficking in urban areas leads to congestion, loss of property, waste of time, damage to the environment, and occasionally to the next level of human mortality. As a result, there is a significant need for smart traffic monitoring and strategies for reduction in cities (Zong et al., 2020). The IoT and ML approaches are the best way to overcome this challenge. It ushers in a new era of intelligent traffic control by effectively aggregating travel times.
This paper aims to utilize IoT, cloud computing, Raspberry Pi, and ML approaches to enhance collection and data analysis. This research is based on existing scenarios: information generated by IoT device data gathered from roads and gates is accessible to all passengers and users (Liu et al., 2020). The system will be able to detect existing traffic, traffic and anticipate future traffic to urban areas when it collects real-time sensor data using unsupervised learning techniques. After that, sensor data monitoring and sensor data detection have been measured for analyzing and visualizing the acquired data. Drivers can use the system-generated data for the optimum route selection. As a result, the system is dynamically administrated, controlling, and monitoring moving cars. The boundary time and conditions of different traveler times may be vague significantly (D. Singh et al., 2017).
The unsupervised learning-based clustering techniques are significantly applied in transport research areas for identifying travel patterns (Pyykonen et al., 2013; Yu et al., 2012). The k-means is a crisp clustering algorithm based on partitioning and the Gaussian Mixture Model (GMM) is a fuzzy-based clustering technique, which is widely used for grouping transport patterns during peak and off-peak hours. To measure the number of clusters necessary for optimal transportation data clustering, this research uses the silhouette coefficient and elbow method.
The rest of this paper is organized as follows: Section 2 shows the related work of this study. Section 3 describes the proposed IoT data acquisition sensor and cluster analysis framework, Section 4 illustrates the outcome evaluation and the outcomes of tests and analysis of the proposed work. Section 5 finally concludes this paper and other improvements are highlighted.