AI-Powered Fare Evasion Detection on Dubai Buses: 90%+ Accuracy Achieved
RTA’s Commitment to Innovation in Public Transport
Dubai’s Roads and Transport Authority (RTA) continues to lead in public transport innovation. Known for operating one of the world’s most efficient and forward-thinking transit networks, the RTA consistently explores technology-driven solutions to improve service quality, optimize costs, and ensure fare compliance.
A Joint Proof of Concept: RTA, CISCO & AWAAIT Target Fare Evasion with AI
In an initiative to address fare evasion — a challenge affecting not only revenue protection but also safety perception and fare fairness — RTA partnered with global tech leader CISCO and Barcelona-based AI specialist AWAAIT. Together, they launched a three-month Proof of Concept (PoC) to evaluate an advanced fare evasion detection system on Dubai’s urban buses.
Mobile App Empowers Inspectors with Real-Time Intelligence
The system was tested on two fully operational buses and leveraged artificial intelligence and machine learning to detect non-paying passengers, with the aim to support enforcement activities in the future— all while minimizing disruption to compliant users.
At the heart of the solution is a dedicated mobile app for fare inspectors, which provides three layers of real-time data:
1. Live Bus Arrival Information
The app shares a good estimate of the number of fare evaders on each bus, allowing inspectors to prioritize inspections based on the number of alleged fare evaders in each incoming bus.
2. Fare Evasion Probability Alerts
The app shares a good estimate of the number of fare evaders on each bus, allowing inspectors to prioritize inspections based on the number of alleged fare evaders in each incoming bus.
3. Visual Assistance for Inspectors
If this layer is activated, the app displays images and short video clips of individuals suspected of not validating their ticket during boarding. This visual support helps inspectors confirm the validity of each alert before focusing their inspection efforts on specific individuals—minimizing inconvenience to compliant passengers. Given the top-down positioning of the 2D cameras mounted on the bus ceiling, identification is based primarily on passengers’ clothing and carried items. For added privacy protection, the system includes an optional blurring feature. Importantly, no facial recognition is used: the final decision to approach a passenger always rests with the inspector, maintaining a human-in-the-loop approach.
Combining Time of Flight Sensors with CCTV for Higher Accuracy
As part of CISCO’s Country Digital Acceleration (CDA) program, AWAAIT was selected for its proven expertise in fare evasion detection using machine vision. A key innovation in the pilot was the use of Time of Flight (ToF) sensors in combination with existing CCTV infrastructure. Unlike video-only systems, with 2D data, ToF sensors delivers depth (volumetric data), hence allowing for a much precise analysis of the boarding scene.
The hybrid system delivered over 90% detection accuracy within a short development timeline. It also provided actionable data by time of day, route, and stop — enabling RTA teams to map fare evasion patterns with unprecedented granularity.
During the PoC, trained observers onboard compared system alerts with real-time observations, confirming reliability. Importantly, inspectors did not intervene; the test was focused entirely on validating detection accuracy in real operational settings.
Clear Results, Scalable Vision
A joint paper by RTA, CISCO, and AWAAIT — presented at the 2024 Global ITS Conference in Dubai — outlined the pilot’s success and its potential for broader application.
Despite strong technical results, the low baseline rate of fare evasion in Dubai meant that large-scale deployment was not cost-justified. However, the pilot provided valuable lessons: AWAAIT is now developing a CCTV-only version of the system, cutting deployment costs and simplifying maintenance.
This simplified version is currently being piloted in a higher-evasion bus network, with an aim to bring AI-powered analytics to the camera edge itself — maximizing scalability, minimizing infrastructure impact, and opening the door to full operational integration.
Discover What AI Can Do for Your Transit Network
At AWAAIT, we believe AI should solve real operational challenges. By working with global leaders like CISCO and forward-looking authorities like Dubai’s RTA, we’ve proven that developing smart, scalable fare control is possible.
Get in touch to learn how we can help your organization improve fare compliance, reduce revenue loss, and boost operational efficiency with AI.