A team of engineering students from a Dubai university has developed an AI-powered system to identify accident hotspots for taxi services in the emirate.
Created by computer science majors at Canadian University Dubai (CUD), the innovative system identifies accident hotspots using data science and AI algorithms and highlights high-demand zones.
The technology aims to reduce accidents and optimise taxi dispatch, ensuring faster customer pick-ups while fostering real-world problem-solving skills among students.
The project, named SmartTranspo, also recently won the prestigious ‘Best Implementation Champion’ award at the Alibaba Cloud/AI and RTA Hackathon 2024 held in Dubai.
Speaking about the project’s inspiration, participating student Mohamad Mourad said, “The hackathon’s theme focused exclusively on solving real problems faced by urban cities, specifically in transportation, with a particular emphasis on Dubai taxis. We were tasked with tackling specific challenges, such as optimising taxi pick-up points, route anomaly detection, smart taxi dispatch, dynamic taxi ranks, accident hotspots, and taxi charging infrastructure.”
Explaining the technical foundation, co-developer Adeeb Mohammad said, “As a team, we were determined to propose an innovative solution using real data. SmartTranspo uses two different training datasets provided by the RTA; one details taxi trips, including trip time, duration, and location, and the other contains data on traffic accidents around Dubai.
“We implemented machine learning algorithms, specifically k-means clustering, to pinpoint exact locations where taxi trips peaked and where traffic incidents were frequent. Using the Google Maps API, we integrated these insights into our application, displaying hotspots in a user-friendly and efficient manner for taxi drivers,” he added.
The students emphasised that safety was their primary focus, aligning their project with the broader goal of reducing accidents in the emirate. According to recent Dubai Police data, 32 people lost their lives in road accidents caused by swerving between January and mid-November. During the first six months of the year alone, there were 262 accidents involving reckless lane changes and sudden turns, resulting in 25 serious injuries and 299 moderate to minor injuries.
“We use AI to analyse real-time data, unlike traditional systems that rely on historical data,” Adeeb explained. “For instance, the accident hotspots at 1pm on a Monday might differ from those at the same time on a Sunday. Our algorithm adapts to these variations using dynamic datasets.”
The team also highlighted the financial implications of accidents for taxi companies. “When a taxi is involved in an accident, the company incurs substantial repair expenses, and the downtime of the vehicle results in lost revenue. Addressing this concern became our priority,” added the third-year engineering student.
The system also identifies high-demand pickup points at specific times using clustering algorithms. “This allows us to optimize pickup locations,” explained Mohamad Hamadeh. “Taxis can be strategically redirected to high-demand zones instead of roaming aimlessly or waiting unnecessarily. As a result, pickup times are reduced, and customer satisfaction improves significantly.”
The team now aims to integrate SmartTranspo with the RTA’s systems, facilitating a broader impact on Dubai’s transportation network.
“This map is designed to seamlessly integrate into various platforms and can be incorporated into Careem, Waze, or Yango’s mapping services,” said Hamadeh. “Looking ahead, we plan to strengthen our technical skills in machine learning, data science, and cloud computing to compete effectively. In the future, we also aim to add features such as taxi charging infrastructure to further enhance the system.”
© Khaleej Times
Twitter