IIIT-H’s AI-powered dashcam system boosts traffic enforcement in Telangana
HYDERABAD: An artificial intelligence (AI)-powered traffic violation detection system developed by the International Institute of Information Technology, Hyderabad (IIIT-H) is helping the Telangana Police improve traffic enforcement by converting patrol vehicles into mobile surveillance units. The system, called VIOLA, detects traffic violations using low-cost dashcams installed in police vehicles and integrates with the state’s existing e-Challan platform.
The technology has been developed by IIIT-H’s Applied AI Research Centre (INAI), building on earlier AI research conducted by the Centre for Visual Information Technology (CVIT).
Govind Krishnan, who leads translational AI initiatives for mobility at INAI, said the objective was not to identify why road accidents occur but to use artificial intelligence to reduce them.
“We already know the answers. The real question is: How can we help solve this problem? Or more importantly, how can artificial intelligence help solve this?” he said.
India accounts for nearly 11% of global road crash fatalities, despite improvements in vehicle safety, stricter traffic laws and greater public awareness. The INAI team believes stronger and more consistent traffic enforcement can significantly improve road safety.
Research evolved into deployable traffic enforcement technology
The project originated from AI-based helmet detection research led by Prof. Ravi Kiran S of IIIT-H’s Centre for Visual Information Technology. INAI expanded the research into a production-ready traffic violation detection system capable of identifying multiple offences.
The work builds on findings from iRASTE (Intelligent Solutions for Road Safety through Technology and Engineering), a road safety initiative involving government and industry partners. The project identified poor road infrastructure and frequent traffic violations—including riding without helmets, triple riding, wrong-side driving and dangerous overtaking—as major contributors to road accidents.
Govind said traffic education remained important, but consistent enforcement proved more effective.
“If a helmet-less rider spots a cop, he’ll immediately stop, pull out the helmet and wear it. But a police officer cannot be present at every junction throughout the day,” he said.
Instead of relying solely on fixed CCTV cameras, the team equipped police patrol vehicles with four dashcams costing about ₹8,000 to ₹10,000 each. The cameras continuously record patrol routes, while AI analyses the footage to identify violations.
The system detects offences such as riding without helmets, triple riding and wrong-side driving. It also uses Automatic Number Plate Recognition (ANPR) to identify vehicle registration numbers before forwarding verified cases to Telangana Police’s e-Challan system.
Prithvi Jonnada, Operations Manager at INAI, said officers continue to verify every AI-generated case before issuing challans.
“This is a mandate from the police department to have a human in the loop, ensuring the AI acts as an assistant, not a replacement, for human judgment,” he said.
System records high accuracy during pilot deployment
According to the INAI team, the four dashcam-equipped patrol vehicles have already contributed to the generation of several traffic challans during the pilot phase.
Prithvi said Telangana Police reported that nearly 96% of violations submitted by the AI system were accurate enough to be approved. He added that the system has become one of the largest contributors to challan generation within the existing e-Challan framework.
“Imagine what happens when this scales to hundreds of vehicles?” he said.
Unlike conventional CCTV systems that monitor only fixed locations, dashcam-equipped patrol vehicles extend surveillance across wider areas without requiring additional roadside infrastructure.
The researchers also said automated detection appears to generate less public resistance than manual enforcement.
“When a police officer uses a handheld device to issue a challan, people often object. But when a CCTV or dashcam automatically detects a violation, there is less confrontation because there is no direct human interaction,” Prithvi said.
The team is now expanding VIOLA’s capabilities to include speed detection and additional traffic offences. The same dashcam footage can also identify damaged roads, potholes and missing lane markings, providing civic agencies with data on road conditions.
The project represents IIIT-H’s effort to translate academic research into deployable public-sector technology. The system received support from iHub-Data, the institute’s technology innovation hub.
The team is in discussions with police departments outside Telangana and is exploring partnerships for wider deployment and commercialisation.
“People don’t expect an academic institution to build a complete industrial-grade product that is deployed and generating real impact. We want to change that,” said Dr Ganesh Yalla, Chief Executive Officer of iHub-Data.
