Coimbatore, July 15 (IANS) More than three months after an AI-powered speed enforcement system was installed on the G.D. Naidu Flyover in Tamil Nadu’s Coimbatore, the project remains in the testing phase, delaying the enforcement of traffic penalties and raising concerns over road safety on one of the city’s busiest elevated corridors.
The 10.1-km flyover, which was opened to ease congestion and improve traffic flow, has witnessed a rise in speeding and other traffic violations since becoming operational. To address the problem, authorities installed 44 artificial intelligence-enabled cameras and 16 digital display boards along the stretch. However, the system has yet to become fully operational because of technical issues and accuracy concerns.
The delay has become a key issue for both traffic police and commuters. During a recent road safety review meeting, police officials urged the State Highways Department to expedite the completion of the trial run, pointing out that the prolonged testing period has limited their ability to effectively regulate traffic on the flyover.
Police officials said the uninterrupted stretch of the flyover encourages motorists to drive well above the prescribed speed limit, increasing the risk of accidents.
While the AI-based surveillance system was introduced to automatically detect violations and reduce the need for manual enforcement, officials said they cannot issue penalties until the software is fully calibrated and approved for regular operation.
Motorists, meanwhile, have expressed concern over the reliability of the system. Several commuters have complained that the cameras occasionally misidentify violations, particularly in detecting seat belt compliance.
Videos highlighting alleged technical glitches have also surfaced on social media, prompting fears among vehicle owners that they could be wrongly penalised despite the system still being under trial.
Traffic police clarified that no challans or fines are being generated during the testing phase and that enforcement will begin only after the technology achieves the required level of accuracy.
Officials from the Highways Department acknowledged the technical challenges and said the private firm responsible for installing and operating the system has been instructed to rectify the defects.
According to the department, the AI software initially recorded an accuracy level of around 20 per cent, which has now improved to nearly 80 per cent through continuous data training and software refinement.
The department has also sought traffic restrictions on parts of the flyover to facilitate further calibration work, as several corrections cannot be carried out while vehicles are moving. Officials expect the system to complete its trial phase and become fully operational within the next month, paving the way for automated traffic enforcement on the flyover.
–IANS
aal/dpb