The issue of noise pollution in urban areas, especially due to vehicle honking, is a growing concern that often remains overlooked. In response, a Judicious Urban Honk Detection Module (JUHDM) has been developed, consisting of two units based on intrusive and non-intrusive principles to accurately identify honking sounds. It is an IoT-based modern world-tested solution. Our research paper[1] delves into the efficacy of these units and combines honking data with a unique driving score mechanism to evaluate driving styles. Concurrently, focusing on road safety, a system named D&RSense is introduced, utilizing smartphone GPS and accelerometers [2]. It categorizes driving behavior, evaluates road quality, and offers real-time alerts through detecting driving events and road anomalies using the Support Vector Machine (SVM) and the Fast Dynamic Time Warping (FastDTW) algorithm. The robustness of this integrated system has been validated through comprehensive experiments, and a novel driver-scoring mechanism is introduced here. Our proposed driver scoring mechanism, which is anchored in empirical observations of real-world driving behaviors, signifies a paradigm shift in the evaluation of driving quality metrics. This system employs a comprehensive and granular assessment methodology, enabling a multi-faceted understanding of driving behaviors rather than a monolithic score. Through its adaptability, it can account for the dynamic nature of driving conditions and user behaviors, thus presenting a forward-thinking approach to the maintenance of driving profiles. Given the critical importance of road safety and the myriad challenges associated with it, the integration of such an advanced scoring mechanism could play a pivotal role in ushering in a new era of enhanced vehicular safety measures in the imminent future.
In this repository, you will get the implementation of the proposed novel driver-scoring Python script along with the dataset on which we have tested the experiment. We are maintaining the repository for the reproducibility purpose of this work for further research. If you are using the code or the dataset, kindly cite our work mentioned below.