Engineering students from a university in Dubai have engineered an Artificial Intelligence solution designed to help track and accurately predict accident hotspots for taxi providers in the emirate. Developed by computer science majors studying at the Canadian University Dubai (CUD), the technology-based system pinpoint the accident-prone zones and areas with high demand.
The technology is to minimize the number of incidents in addition to help in accomplishing enhanced delivery of taxi services to the customers as well as enhancing practical problem solving skills among the students. SmartTranspo project, the solution recently clinched the Alibaba Cloud/AI and RTA Hackathon 2024 competition’s ‘Best Implementation Champion’ award.
Table of Contents
Tackling real-world challenges
When asked about the inspiration behind the project, one of the participating students, Mohamad Mourad shared that the hackathon’s topic was set around finding solutions to the actual challenges of the urban cities concerning transportation with the main emphasis on the Dubai taxis. Our focus was in addressing certain issues, like optimal taxi stand locations, changes in route patterns, efficient dispatching of taxis, dynamic standing places of taxis, accident prone areas and charging infrastructure of the taxi.
Using data-driven insights
The technical rationale was clarified by co-developer Adeeb Mohammad as the team aimed, “To put forward something new and practical based on actual data.” SmartTranspo employs two types of training datasets from the RTA; one comes with the information on taxi travel in terms of time, duration and location while the other gives information on traffic accidents report in Dubai.
We used k-means clustering algorithm to identify areas that registration spiked at, and where taxi trips happened simultaneously with traffic issues. Specifically, these insights were implemented into the application by means of the Google Maps API; the interface with the hotspots is comprehensible and convenient for the use by the representatives of the taxi service, he further explained.
Addressing safety concerns
Responding to the questions about their project, students stressed that the task for them was to increase safety and to prevent the further rising of accidents in the emirate. The latest statistics from Dubai Police revealed that 32 people were killed in road accidents due to fluctuation between the inner and outer lanes within seven months from January to mid-November. Even in the first six months of the year, 262 road accident cases were reported due to reckless lane change and sudden turn, which caused 25 major and 299 minor and moderate injuries.
Also Read: Indians Facing High Rejection Rates for Dubai Tourist Visas Under New Rules
“We actually deploy AI to work real time to try and parse data, where traditional systems work on history,” Adeeb said. For example, the accident-prone areas in the middle of a working week, namely at 1 pm on Monday may not be the same as at 1 pm on Sunday. Our algorithm adjusts the variations with dynamic data sets.”
The team also pointed that accidents also have monetary consequences for the owners of taxi cars. “Whenever a taxi has been involved in an accident the costs for repairing such a car is very expensive and apart from that, the time that the car must remain idle is also an added cost to the company. Addressing this concern became our priority,“ said a third-year engineering student.
Optimising taxi services
The system also uses clustering algorithms to determine high demand pickup points at certain times. “It helps us to address issues such as the choice of pickup locations, according to Mohamad Hamadeh. Taxis can be directed to specific areas where they are likely to be needed more than cruising around or idling. Consequently, available time for pickups is cut down, and customer satisfaction is enhanced a notch up.”
The team now wants to integrate SmartTranspo with the RTA’s systems to have a more extensive influence on transportation in Dubai.
“This map can be easily embedded into the platforms and one can even add it into the Careem or Waze’s or Yango’s navigation service,” concluded Hamadeh. “More specifically for the future our key tactic is to develop an even stronger technical competency based on the areas such as machine learning, data science and cloud computing as a way to fairly compete. In the future we also want to incorporate such sub-functions as charging infrastructure for taxis to optimize the system.