Artificial Intelligence (AI)-powered solutions may soon make Indian roads safer to drive on. The Indian government announced on Tuesday that artificial intelligence-powered technology could reduce the risk of road accidents in the country, which killed over a lakh people in 2020.
To avoid this, the Indian government announced that the AI approach will employ a first-of-its-kind dataset of 10,000 images.
This dataset is finely annotated, with 34 classes collected from 182 drive sequences on Indian roads obtained from a front-facing camera attached to a car driving around the cities of Hyderabad and Bangalore and their outskirts, according to the company.
According to India’s Ministry of Science and Technology, this novel approach uses the predictive power of AI to identify road risks and a collision alert system to communicate timely alerts to drivers in order to improve road safety.
It is currently being implemented in Nagpur, Maharashtra, and the government intends to roll it out throughout India.
The project’s goal is to reduce traffic accidents. According to WHO, road accidents kill up to 1.35 million people worldwide each year.
The Nagpur-based project ‘Intelligent Solutions for Road Safety through Technology and Engineering’ (iRASTE) will identify potential accident-causing scenarios while driving a vehicle and alert drivers about them using the Advance Driver Assistance System (ADAS).
The project will identify ‘grey spots,’ i.e., through data analysis and mobility analysis while continuously monitoring dynamic risks across the entire road network, the Indian government stated.
Grey spots are areas on roads that, if not addressed, can become blackspots (locations with fatal accidents).
According to government officials, the system also performs continuous road monitoring and designs engineering fixes to correct existing road blackspots for preventive maintenance and improved road infrastructure.
The iRASTE project is supported by the Department of Science and Technology and is managed by the I-Hub Foundation at IIIT Hyderabad, a Technology Innovation Hub established in the technology vertical of Data Banks and Data Services (DST).
For several other data-driven technological solutions in the mobility sector, the I-Hub Foundation has used techniques such as machine learning, computer vision, and computational sensing.
The India Driving Dataset (IDD) is one such solution. It is a dataset for road scene understanding in unstructured environments captured from Indian roads that deviate from global assumptions of well-defined infrastructure such as lanes, limited traffic participants, low variation in an object or background appearance, and strict adherence to traffic rules.
The project consortium includes CSIR-CRRI and Nagpur Municipal Corporation, as well as industry partners Mahindra and Intel.
The Hub’s mission is to coordinate, integrate, and expand basic and applied research in broad data-driven technologies, as well as to disseminate and translate it across the country.
One of the main goals is to create a critical resource for future use by researchers, startups, and industry, primarily in the areas of smart mobility, healthcare, and smart buildings.
What makes the iRASTE project even more unique is that AI and technology are being used to create practical solutions for Indian conditions as a blueprint. While iRASTE is currently being implemented in Nagpur, the ultimate goal is to replicate the solution in other cities as well.
The Telangana government is currently in talks about incorporating the technology into a fleet of highway buses. There are plans to expand the scope of iRASTE to Goa and Gujarat as well according to ministry officials.
Another dataset, Open World Object Detection on Road Scenes (ORDER), was created using the India Driving Dataset and could be used by autonomous navigation systems in Indian driving conditions for object localization and classification.
A Mobility Car Data Platform (MCDP) has also been designed with several sensors – cameras, and LIDARs, with the necessary computing for anyone to capture or process data on the car that can help Indian researchers and start-ups test their automotive algorithms and approaches in navigation and research on Indian roads, according to the Indian ministry.
LaneRoadNet (LRNet), a new framework with an integrated mechanism using deep learning to consider lane and road parameters, has been designed to address problems on Indian roads, which have several obstacles, such as occluded lane markings, broken dividers, cracks, potholes, and so on, that put drivers at significant risk while driving, according to the government.
A road quality score is calculated in this framework using a modular scoring function. The final score assists authorities in assessing road quality and prioritizing road maintenance schedules in order to improve drivability.
