Digitally enabled and intelligent technologies are finding their ways in manging our transport systems in a more sustainable manner. These technologies include but are not limited to data collection using various sensing systems and internet of things (IoT) concepts as well as using smart phones and other digital technologies while leveraging machine learning (ML) algorithms and artificial intelligence for analysing and interpreting the data. These are then used, or can be used, as part of an automated asset-management system. To this end, both the transport industry and academia have produced some innovative and exciting solutions and findings, some of which are showcased in this themed issue. The issue concentrates on digitally enabled and intelligent technologies for a sustainable transport, a topic which is briefly described here.
Sustainable transport is a broad and sometimes debatable topic, but, considering the ever-growing urban population, improving air quality and achieving low carbon dioxide (hereafter ‘low-carbon’) or net-zero transport have been the focus of many national and local authorities as part of a sustainable development goal. To this end, Peng et al. (2026) investigated the key contributors of urban low carbon travel levels and analysed the associated impact mechanisms using mobility data extracted from the mobile phone network in Wuhan, China. ML algorithms were used to analyse the data and to assess the low-carbon travel levels at a 1 km2 scale. Land use density and diversity were identified as the main two factors followed by building density, surface roughness and public transportation provision. While the study was based on some assumptions to be able to interpret mobile phone data, it shows the application of such a data to determine mode of travel, while the ML algorithms show promising results in identifying the most significant factors. Similarly, Ergin (2026) investigated the application of global and station-based extreme gradient boosting (XGBoost) methods – ML algorithms – in predicting urban air quality. Focusing on spatial homogeneity, the methods were tested on data from Istanbul, Türkiye. The global method showed promising results for coarse particulate matter (PM10; R2 = 0.92) but failed to accurately predict nitrogen dioxide (NO2; R2 = 0.35), while the station-based method achieved an R2 of 0.93 for nitrogen dioxide. The findings showed the importance of selecting the right ML models to avoid unacceptable results. A similar observation was made by Aksoy (2026), who adopted an artificial neural network method for forecasting the total travel time. The adopted method was able to reduce the computation time by 60 000 times compared to the traditional traffic-assignment model.
Improving transport safety and reducing road and railway accidents have been another area of focus for many authorities around the world, considering the huge negative impacts of accidents to the affected society and individuals. Kuşkapan and Çodur (2026) investigated the application of ML models for safety of pedestrians, one of the most vulnerable groups. The study used the models to assess pedestrian crossings (a total of 719) in Erzurum, Türkiye, which showed 74% of the assessed pedestrian crossings have safety issues, detected by a selected ML algorithm. Similarly, Özkan et al. (2026) investigated possible enhancements of railway crossing safety using an ML model to forecast accidents at a crossing and to select the most cost-effective safety measures using data from 597 crossings in Türkiye (2015–2023). The adopted random forest model was reported to have 88% accuracy in identifying the dominant risk factors.
Another investigated application of digital technologies is on asset inventory and condition assessment of various types of transport systems. For example, Lei et al. (2026) used image-processing techniques for localising asphalt pavement cracks using limited annotated data sets, compared to the large data sets required by deep learning models, but with higher accuracy than traditional image processing methods. The proposed methodology was implemented on an asphalt-pavement repair robot with a depth camera, which showed over 90% improvement in accuracy compared to the traditional method.
Implementation of digital technologies, despite their promising results, can impose challenges to an organisation, such as the one identified by Al Suwaidi et al. (2026) who focused on artificial intelligence implementation in transportation and the associated challenges in United Arab Emirates. Using interviews and questionnaires from around 400 participants at managerial level, the study reported required capital costs, lack of essential infrastructure, existence of high-quality data and funding/resources as the most significant barriers in the country.
While various studies included in this themed issue draw a promising future for a digitally enabled and intelligent sustainable transport, there are still areas for further research in the field, including development of frameworks to assess the maturity level of existing digital infrastructure and to be able to suggest a cost-effective pathway towards a higher level of sustainable transport for any transport authority around the globe.
