Efficacy of Bluetooth-Based Data Collection for Road Traffic Analysis and Visualization Using Big Data Analytics

Abstract

Effective management of daily road traffic is a huge challenge for traffic personnel. Urban traffic management has come a long way from manual control to artificial intelligence techniques. Still real-time adaptive traffic control is an unfulfilled dream due to lack of low cost and easy to install traffic sensor with real-time communication capability. With increasing number of on-board Bluetooth devices in new generation automobiles, these devices can act as sensors to convey the traffic information indirectly. This paper presents the efficacy of road-side Bluetooth scanners for traffic data collection and big-data analytics to process the collected data to extract traffic parameters. Extracted information and analysis are presented through visualizations and tables. All data analytics and visualizations are carried out off-line in R Studio environment. Reliability aspects of the collected and processed data are also investigated. Higher speed of traffic in one direction owing to the geometry of the road is also established through data analysis. Increased penetration of smart phones and fitness bands in day to day use is also established through the device type of the data collected. The results of this work can be used for regular data collection compared to the traditional road surveys carried out annually or bi-annually. It is also found that compared to previous studies published in the literature, the device penetration rate and sample size found in this study are quite high and very encouraging. This is a novel work in literature, which would be quite useful for effective road traffic management in future.

Data-Driven Traffic Assignment Through Density-Based Road-Specific Congestion Function Estimation

“…The limitations of the loop detectors of MIDAS reduce the accuracy of the topographic representation, limiting the fidelity of the traffic assignment results. Improvements to this could be made by augmenting the coverage of the loop detectors with other data, such as from roadside Bluetooth data [45]. Also, if such data sources can differentiate measurements for different vehicle types, this could allow BPR-Density to be applied to multiclass traffic (i.e.…”

Real-Time Monitoring of Data Pipelines: Exploring and Experimentally Proving that the Continuous Monitoring in Data Pipelines Reduces Cost and Elevates Quality

“…By doing so, we can ensure that stakeholders receive the right information at the right time, enabling them to make timely decisions. Additionally, adopting an effective monitoring and alerting system can result in significant cost savings and an enhanced end-user experience [4]. To address the unique challenges posed by Data Applications, a dedicated monitoring framework must be established for Data pipelines.…”

A Review of Data Mining, Big Data Analytics and Machine Learning Approaches

“…The challenges associated with deep learning in the realm of Big Data analytics refers to the incremental training for dataset that are non-stationary, large-scale frameworks, and high-dimensional data, as indicated by Kulkarni, Kumar, and Rao [33]. The Variety feature of the big data analytics refers to the input data diversity of its domains and types within the realm of big data.…”