Dubai’s Roads and Transport Authority introduced a new Smart Test System for light-vehicle yard tests on 29 September 2026, with software calculating results against standardised criteria without human intervention, Gulf News reported. The system combines artificial intelligence, computer vision, vehicle positioning and data analysis to assess learners as they carry out manoeuvres in the testing yard.
Key points
- The Smart Test System calculates light-vehicle yard test results without a person intervening in the calculation.
- RTA says its positioning technology tracks vehicles in the yard with accuracy of up to two centimetres.
- Learners can review their route and recorded errors on an interactive 3D map after a test.
- Facial recognition checks identity, while an automatic braking system can intervene when it detects a collision risk.
RTA assigns yard test calculations to software
The change concerns how a result is calculated in a light-vehicle yard test. The system monitors a learner’s performance in real time and assesses manoeuvres against standardised criteria. It also records behaviour beyond the vehicle’s path: cameras check for mirror and blind-spot checks and monitor where the learner places their hands on the steering wheel during manoeuvres.
Sultan Al Akraf, Director of Drivers Licensing at RTA’s Licensing Agency, said the authority wanted learners assessed against consistent criteria. He said the system was intended to reduce differences in assessment and give customers a clearer account of their results. The cameras, vehicle tracking and automated calculation are parts of that assessment, rather than separate tests presented to the learner.
Facial recognition verifies who is taking the test so that the result is associated with that learner. Khaleej Times reported that the system uses the identity check alongside its real-time assessment of driving manoeuvres. The automated calculation applies to the result of the yard test; the authority’s description of the system concerns that testing setting.
Two-centimetre positioning records yard manoeuvres
RTA says the upgraded system uses Real-Time Kinematic positioning to locate vehicles within the testing yard with accuracy of up to two centimetres, Gulf News reported. The location data lets the system identify mistakes involving a vehicle’s route or parking position. Computer vision supplies information about the learner’s actions inside the vehicle, including checks of mirrors and blind spots.
Those observations feed an assessment made while the test is taking place. Dubai Eye reported that artificial intelligence, computer vision and data analytics monitor manoeuvres and safety-related behaviour in real time, with results calculated automatically against standardised criteria. Vehicle position and driver behaviour therefore enter the same assessment, rather than leaving the result to be calculated by an examiner after the drive.
RTA says the equipment is designed to operate consistently across different weather and lighting conditions, including tests held during the day and at night. That design is intended to limit the effect of external conditions on assessments. Al Akraf said consistent evaluation was one aim of the upgrade, alongside giving learners a clearer account of how their performance was judged.
The test vehicles also have preventive automatic braking. The system can apply the brakes when it detects circumstances that could lead to a collision, including a risk to the learner, the vehicle or the surrounding area. Al Akraf said the collision-avoidance and emergency-braking functions were intended to make testing safer by addressing risks before an accident occurs.
A 3D map for Dubai learners
After the test, learners can replay the vehicle’s journey and inspect an interactive 3D map marking where errors occurred. The visual evidence identifies major and minor mistakes, giving a learner a way to review the route as well as the result. Khaleej Times reported that the system is intended to help learners understand which skills need further work before another test.
Al Akraf described the visual evidence as part of a wider account of performance across manoeuvres. The system presents recorded errors to customers so they can examine the reasons for a result and identify areas to improve. For a learner, that means the account of a failed manoeuvre can include where it happened on the route, rather than only an entry in a list of errors.
Performance data can also be integrated with the driver training system. Instructors could use information about an individual learner’s errors to identify areas needing more practice and tailor lessons accordingly. Such integration could let instructors use the learner’s performance to focus subsequent training.
