Acoustic Sensors and Audio Signal Processing in Intelligent Transportation Systems: A Survey
Hossein Parineh, Majid Sarvi, Saeed Asadi Bagloee
Abstract
The effectiveness of intelligent transportation systems (ITS) strongly relies on the quality of input data provided by traffic monitoring sensors. Conventional traffic sensors such as cameras and radar necessitate costly infrastructure and are limited in the types of data they can collect. In contrast, acoustic sensors emerge as an affordable alternative that provides a wide range of road-related data (e.g. vehicle count, speed estimation, etc.). This paper presents the first survey encompassing the applications of audio signal processing (ASP) and acoustic sensors within the domain of ITS. Through an extensive literature review, this study examines various tasks relevant to traffic sensors, characterizes frequent feature extraction methodologies, and evaluates ASP models. Moreover, the paper investigates recent trends in AI-based ASP applied in ITS, which have resulted in more sophisticated acoustic sensors. To furnish a pragmatic roadmap for future investigations in this domain, this survey conducts a comparative analysis of studies in the same class, categorized by each task assigned to acoustic traffic sensors, and illuminates their strengths and weaknesses. Lastly, the paper identifies the main research gaps and suggests future directions aimed at advancing ASP in ITS. In conclusion, this study provides insights and recommendations to guide future research on this subject and serves as a reference to the state-of-the-art of ASP for acoustic sensors deployed in ITS.