Integrative review of data sciences for driving smart mobility in intelligent transportation systems

As intelligent vehicles (IVs) continue to advance in fully connected environments, the collection of data from various sources in intelligent transportation systems (ITSs) has reached unprecedented levels. This paper aims to provide an integrative review of the processing and utilization of this vast data for optimizing smart mobility (SM) and extracting actionable insights to enhance planning and decision-making. While the data science (DS) frameworks have proven its effectiveness in sectors such as healthcare, tourism, social media, and the internet industries, there remains a lack of systematic research on DS in the context of SM (referred to as DS²M) within the ITS field. In this paper, we examine the potential applications of DS in IV systems by exploring relevant literature in DS domains, including discussions on data uncertainty, deep learning-based interpretability, reinforcement learning, and the relationships within IV data. These applications include IV control systems, data analytics visualisation, parallel-driving IV systems, and other DS²M applications. Furthermore, the analysis of seminal and recent literature emphasizes the absence of widely recognized benchmarks, which poses challenges to the validation and demonstration of new studies in this evolving domain.