Artificial Intelligence in Road Safety

Artificial intelligence (AI) is the emulation of human intelligence in computers that have been programmed to think and act similarly to humans. A humanoid is a machine that has human-like characteristics including learning and problem-solving.

Despite various hopeful initiatives to reduce road accidents, they continue to claim the lives of millions of people each year. Over 1.35 million people are killed in automobile accidents every year, with another 20-50 million suffering non-fatal injuries (WHO, 2020). Around 80% of all accidents are caused by human error, such as inattention, over-speeding, or an unsuitable lookout. Various safety precautions, such as airbags, seatbelts, and speed limits, are now in place to lessen the number and severity of accidents.

The reason for unsafe roads is complex and multi-faced, requiring a systematic approach: road safety management, safer road users, safer vehicles, safer infrastructure, and post-accident care. The conventional measures of road safety are not yielding the desired results as evident from the fact in the past decade over 13 lakhs people lost their lives on Indian roads.

This is time to upgrade conventional road safety to technology-driven Road Safety as?despite various initiatives it continues to claim lives as over 150,000 people succumb to road fatalities causing an economic dent of over INR 4 trillion to GDP apart from unimaginable miseries to the families of road accidents victims.

Artificial intelligence (AI), deep learning, and IoT can play a life-saving role in this digital age. AI will improve as computing power as well as unsupervised learning and pattern recognition. In Tesla's Model 3, three AI-based applications, including obstacle, traffic sign, and cut-in recognition, explain AI's potential for road safety. In particular, the application of convolutional neural networks for image processing gives good hope.

The 4 E's of Road Safety

The 4 E's emphasis on accident prevention and control are (i) Education, (ii) Enforcement, (iii) Engineering, and (iv) Emergency care. There are plenty of scopes to introduce AI in all the 4 Es of road safety to take it to the next levels.

Education:

AI can change the landscape right from the issuance of driving licences to driver’s education and ascertaining regular skill levels using other cofactors and then recommend for specific up-skilling. AI can easily detect drowsiness and give alerts to prevent accidents, thus compensating for human limitations. Basic ML algorithms can be used to track driving behaviour and convert driving behaviour into good driving scores (GDS). The GDS can be used for incentivising good driving behaviour even when no one is watching. The higher the GDS, the better is the reward. The reward can be in form of rebate in Insurance premiums, rebate in fuel, rebate at vehicle repair garages/workshops, tyres purchase and even at wayside amenities.

AI can also determine the spot speed limit based on localised traffic, road conditions and weather conditions and assist the drivers with real-time scenarios.

Enforcement:

Nearly 15 million reported traffic violations take place annually, these violations lead to collisions and accidents. Manually it is impossible to detect such a large number of traffic violations and penalise them. As a result, a lot of violators are never caught and penalised, encouraging others.?AI can assist the enforcement system in the early detection of traffic violations using AI, IoT, and Speed cameras and generating real-time automated penalty/fine slips for traffic violations.

Though the Motor Vehicles Act of 2019 and the Central Motor Vehicles Rules include several rules and the state governments are largely responsible for enforcing these regulations but there has been a gap connecting the dots, which can be filled by AI-based platforms.

Engineering:

AI can be of great assistance in accident investigation, road safety management, road safety audit, identification of black spots (accident-prone areas). Identifications of dangerous potholes or potential landslides in hilly roads can be predicted by AI better.

Emergency Care:

Post-accident emergency care is another serious concern in India as nearly 50% of people die on the spot or while transiting to the hospital. If the occurrence of accidents and location can be shared to provide emergency care within golden hours a lot of lives can be saved. AI can identify the first level visual injuries by image processing and assist the paramedics on the spot with better medical advice.

?WEF led Road Safety 2.0 is one such next-generation approach, which focuses on the infusion of Artificial Intelligence in road safety to achieve quicker results, to scale up awareness of the socio-economic impact of road accidents and to assist in creating a financially self-sustainable ecosystem in public-private partnership model towards road safety in India.?