Research News

Deep Learning Approaches for Vulnerable Road Users’ Safety: Special Journal Issue Edited by Khalifa University Expert

February 15, 2023

Khalifa University’s Prof. Ernesto Damiani was invited to serve as a guest editor of a special issue of Safety Science dedicated to the latest technological developments and innovations that will improve safety in dangerous traffic situations. 


According to the World Health Organization’s Global Status Report on Road Safety 2018, the number of road-traffic deaths continues to climb with approximately 1.35 million people dying every year in road accidents. Of these, the burden of road-traffic injuries and deaths is disproportionately borne by vulnerable road users who contribute to half of all victims. In other words, almost 50 percent of road fatalities are among cyclists, pedestrians or other people who use the roads but aren’t using motorized vehicles.


While progress has been made by strengthening legislation and encouraging the use of seat belts, less focus has been placed on the planning, design and operation of roads and roadsides, which can be assisted by deep-learning approaches.


Deep learning is a machine-learning method based on neural-network architectures with multiple processing layers— it’s a technique that teaches systems to learn by example. Deep-learning algorithms extract patterns from data and learn to associate these patterns with future data. This ability to learn from experience has made deep learning widely applied in various application areas, including health care, visual recognition, text analytics and cybersecurity. However, building an appropriate deep learning model is a challenging task.


Safety Science dedicated a special issue to deep-learning approaches for vulnerable road users’ safety and asked experts from the journal’s editorial board to serve as guest editors. Prof. Ernesto Damiani, Director of the Khalifa University Center for Cyber Physical Systems, was invited to edit this edition, alongside three other editors from Korea, Italy and Canada.


“Deep-learning methods with the benefits of improving accuracy and enhancing efficiency have been broadly adopted in both industry and academia,” Prof. Damiani said. “From a transportation point of view, these methods bring about both challenges and opportunities. They can help improve safety in dangerous traffic situations while also being used for novel transportation applications, including autonomous vehicles.”


The editors selected articles focusing on state-of-the-art theories and novel-application scenarios, surveys of recent progress in the field, and novel-analysis methods and applications. They also featured articles on building benchmark datasets.


The articles covered applications for deep-learning approaches, including data collection on who would be considered a vulnerable road user, intelligent cooperative traffic systems, road-traffic measurements and modeling, and cybersecurity countermeasures to protect systems and prevent dangerous behaviors.


“The articles in this special issue provide insights in a field related to safety science using deep learning and intelligent edge computing, including models, performance evaluation and improvements, and application developments,” Prof. Damiani said. “We hope readers can benefit from these insights and contribute to these rapidly growing areas. We also hope this issue encourages the scientific community to pursue further investigations leading to the rapid implementation of these technologies and approaches.”


Jade Sterling
Science Writer
15 February 2023