Last month, a bus carrying 42 passengers in Almora district of Uttarakhand “skidded before tumbling down a 60-metre gorge”.

Barring up to six survivors, all the other passengers died in the crash.

Road mishaps, whether on India’s highways or city streets, happen with such regularity that they do not stir either the central or state governments into taking strong preventive measures.

But advances in artificial intelligence could offer new solutions to tackling India’s road toll.

India’s road trauma burden

While over the last two decades there have been significant investments in the development of better road infrastructure in India, challenges persist.

These are in the form of poor maintenance of roads, inadequate safety measures, urban congestion, reckless driving, insufficient protection for vulnerable road users, weak enforcement of traffic laws and a lack of education and public awareness about safe driving practices.

As a consequence, India ranks as the top country in terms of the global road fatality rate, accounting for 11.7% of all deaths.

In 2022, India reported over 460,000 crashes that resulted in 168,491 deaths and 443,366 injuries.

The previous year, the numbers stood at 412,432 crashes, 153,972 deaths and 384,448 injuries.

This marks an increase of 11.9% for accidents, 9.4% for deaths and 15.3% for injuries.

While Tamil Nadu recorded the most number of road accidents between 2018 and 2022, the most deaths during the same period were in Uttar Pradesh where 22,595 and 21,227 people died in road crashes in 2022 and 2021.

Over speeding is often attributed as the main cause for road deaths in Uttar Pradesh.

How can we reduce the number of accidents on India’s roads?

To get an overall sense of the reasons for road accidents, the resulting deaths and what could be done to reduce them, the Central Road Research Institute analysed data related to Nagpur, in Maharashtra, between 2008 and 2021.

With a total population of 3.06 million people, Nagpur witnessed an annual average of 200 deaths and about 1,000 injuries, which is on the higher side for Maharashtra.

This study highlighted the urgent need for marrying engineering solutions with technology to address urban traffic problems such as risk of collisions.

The AI-based technology – advanced driver assistance systems – involves mounting a camera on a vehicle’s windshield to scan the road ahead and uses complex algorithms to track potential risks.

In the event of a potential collision, the systems send an audio and a visual warning to the driver. A similar warning is sounded when collision situations involving pedestrians, cyclists or stray animals arise.

With India’s central government’s avowed goal to achieve a 50% reduction in road deaths and injuries by 2030, one of the measures open to it is adopting such AI-based systems.

These systems use road-facing cameras to capture images or video to detect static and moving objects within a predefined distance and time range and identify the lane used by a vehicle in motion.

Such systems can then alert drivers in real time if they are at imminent risk of crashing.

In addition to alerting drivers, the systems collect timestamped details of generated alerts which include the identity of the driver and their geographical coordinates (latitude and longitude), allowing real-time mapping of potential crash risk areas or grey spots.

Nagpur as a test case

The analysis of the data led to a pilot project dubbed iRASTE – or Intelligence Solutions for Road Safety through Technology and Engineering – which was launched in Nagpur in September 2021.

This initiative aimed to reduce road crashes and deaths in the Nagpur metropolitan area by using public buses to collect real-time data of what was happening on the city’s roads.

Cameras were installed in 150 buses which operated in Nagpur’s urban and peri-urban road network, providing various visual and audio alerts at a minimum of 2.5 seconds before potential accident situations arose.

These alerts included forward collision, headway monitoring or the distance between you and the vehicle directly in front of you, lane departure and pedestrian and cyclist collision warnings.

The alerts were deemed to be effective in helping drivers make safer decisions on the road.

The main thrust of road crash prevention relies on the four Es of road safety – education, engineering, enforcement and emergency care.

This pilot project was designed to address four factors – vehicle safety, infrastructure safety, mobility safety analysis and education and awareness and emergency care – which all align with these four Es.

At a later stage, the project introduced a fifth E – encouragement – which took the form of incentivising bus drivers to better maintain their buses and practice safe driving behaviours.

Reduced road crashes

One of the pilot project’s key findings was that these driver interventions potentially reduced crashes by 60% to 66% and deaths by up to 40%.

Thirty-six lives were saved as a result of timely interventions between July 2023 and April 2024 at Nagpur’s grey spots.

As part of the pilot project, RFID-based scanners installed at 10 traffic intersections (black and grey spots) resulted in a 24% improvement in signal adherence among bus drivers in December 2023.

While black spots were identified as intersections where most mishaps occur, grey spots included traffic signals where accidents were comparatively fewer.

A before and after assessment for the periods January 2021 to September 2022 and October 2022 to December 2023, revealed that the installation of cameras in public buses resulted in the reduction of road crashes by 33% and deaths dropped by 100% during the study period.

However, the number of injuries remained constant. This suggests that while such systems can reduce the occurrence and severity of crashes, other measures are needed to prevent injuries.

Additionally, it was found that there was a complete reduction in bus crashes across two segments of the city, with zero recorded incidents after the cameras were installed.

However, crashes went up by 100% and injuries by 75% in the third zone. This suggests variations in driver training, route conditions and other external factors were also factors.

Pilot projects such as these show that AI-based tools have the potential to significantly reduce road crashes and fatalities, and hence improve road safety.

This initiative illustrated the impact of technology in high-risk areas, paving the way for broader adoption across varying types of vehicles.

More importantly, similar systems and technology should be used in India’s megacities and highways where road accidents go unchecked.

S Velmurugan is Chief Scientist and Head, Traffic Engineering and Safety Division, Central Road Research Institute, New Delhi.

Additional reporting and contributions to this article by Dev Singh Thakur, a researcher at Advanced Mobility Analytics Group in Hyderabad and former Research Engineer at the International Institute of Information Technology in Hyderabad.

Originally published under Creative Commons by 360info™.