Introduction

The application of AI in urban transport, from intelligent traffic management to autonomous vehicles and decision-making systems, has significant potential to reduce emissions, optimize energy consumption and improve citizens’ access to transport services. However, despite the rapid expansion of this field, collective knowledge about its conceptual structure, evolutionary trends and future directions is still scattered and unorganized. This is doubly important because AI can act as a tool for achieving sustainable development goals such as sustainable cities and climate action, and on the other hand, if designed inequitably, it can exacerbate social and accessibility gaps. In this regard, the main question of the present study is: How has the global knowledge structure in the field of AI in sustainable urban transport been formed, what trends and dominant conceptual clusters exist in it and what research and policy gaps does it face? Previous studies have mainly focused on technical or case-specific aspects and have not addressed systematic and comprehensive analysis of global literature with a scientometric approach, especially identifying research gaps and future directions. Therefore, the present study aims to fill this gap by systematically mapping the knowledge of this field over the period 2005 to 2024. Based on this, the main objectives of the present study are: identifying temporal and geographical trends in knowledge production; analyzing conceptual clusters and keyword co-occurrence networks; identifying research gaps and providing future directions.



Methodology

The present study is descriptive-analytical in terms of method with a qualitative-quantitative scientometric approach; because in addition to describing the patterns of document analysis, keyword co-occurrence, overlap and vocabulary density, it has been attempted to identify conceptual clusters and semantic relationships of conceptual clusters of artificial intelligence in sustainable urban transportation, as well as identifying research gaps and future directions. Also, this study is considered an applied research in terms of purpose. The statistical population includes all sources indexed in the Scopus database from 2005 to 2024, which includes 172 sources. In fact, the basis of the present study is research that has been conducted in the last 20 years. To retrieve documents, a structured search was performed with the aim of covering the broadest possible meanings of two key concepts: TITLE-ABS-KEY ( "artificial intelligence" OR "AI" OR "machine learning" OR "deep learning" OR "neural network" OR "intelligent system") AND (TITLE-ABS-KEY ( "sustainable urban transport" OR "sustainable mobility" OR green transport" OR "eco-friendly urban mobility" OR "low-carbon urban transport")). The reason for choosing Scopus is that this scientific database has been analyzing data in the VOSviewer software, whose important advantage is processing a huge volume of data and drawing a visual network map of them.





Results and discussion

The most important findings of the study show that the production of scientific resources in the field of AI applications in sustainable urban transport has gone from gradual growth (2005–2015) to accelerated growth (from 2021 onwards). This acceleration is influenced by four key factors: technological developments in AI (especially deep learning and natural language processing from 2018 onwards), global political and environmental pressures (such as the Paris Agreement and the Sustainable Development Goals), the expansion of smart cities, and the impacts of the COVID-19 pandemic. Geographical analysis reveals the prominence of countries such as India (with 33 sources), China (21 sources), the United States, Germany, and Italy, which is due to a combination of rapid urbanization, supportive government policies, and strong research infrastructure. Co-occurrence analysis of terms reveals three conceptual layers: technology (AI, big data), operational application (urban transport, traffic), and policy (sustainable development, urban planning). Additionally, six conceptual clusters are identified that demonstrate the semantic relationship between technology, sustainability, user behavior, policy, optimization, and spatial decision-making. However, research gaps are evident, such as the lack of attention to social justice in mobility, the neglect of algorithmic aspects and emerging technologies (such as autonomous vehicles), and the lack of comparative, context-based studies. Future directions require a focus on AI for justice, the integration of environmental indicators into algorithms, the development of context-based solutions for developing cities, and deeper interdisciplinary research.



Conclusion

The research findings indicate that the field of AI in sustainable urban transport has undergone extensive changes, as it has transformed from a novel idea in the first decade of this century to a dynamic and multidimensional research trend with rapid growth since 2021. Factors such as digital advances, environmental crises and global climate pressures, the development of smart cities, and developments after the COVID-19 pandemic have made this field a dynamic and multifaceted topic. In addition, the geographical distribution of resources indicates the leadership of countries such as India, China, and the United States, reflecting the effects of rapid urbanization and national policies on knowledge production. Based on scientometric analyses, six key conceptual clusters in the field of AI in sustainable urban transport have been revealed. These six conceptual clusters include AI technologies, environmental and sustainability components, user behavior and mobility, transportation policy, system optimization, and decision support systems. At the same time, emerging concepts, including terms such as smart cities, deep learning, and micromobility, have recently attracted serious attention from researchers and policymakers. However, the social and algorithmic aspects have received less attention.