The Telangana AI Mission (T-AIM) recently launched the Mobility AI Grand Challenge, seeking artificial intelligence-enabled solutions to detect and classify potholes based on their severity in the city of Hyderabad. T-AIM launched the challenge on behalf of the local government body, the Greater Hyderabad Municipal Corporation (GHMC). The solutions will help GHMC undertake targeted repair work projects in the city.
Using AI, participating teams will need to analyse video feeds (40 fps or higher) from specific routes in Hyderabad for potholes. The potholes in the video feed need to be identified and classified using AI. The classification of pothole severity may be undertaken based on the surface area and depth of the pothole, the official T-AIM portal mentioned.
Apart from the opportunity to win financial support to the tune of IN ₹2 million (US$ 25,118) to implement pilot projects, start-ups will be able to network with AI leaders, government officials, and industry players. It provides a platform for new innovators to showcase their AI solution development capabilities at a larger scale.
T-AIM has outlined the criteria for start-ups to participate. Indian startups registered anywhere in India can participate and they should be in existence for no older than eight years (incorporated not prior to 1 January 2014). Innovators will be selected based on their approach note and the shortlisted ones will get six weeks to present a proof-of-concept.
The teams will be given four weeks to work on the use case and will then have to submit the results, approach, and code for evaluation by the jury. They will also have to host the solution online to be assessed. The winning criteria will be a combination of quantitative and qualitative methods to solve the problem. The first criterion is the Approach, which judges the team’s description of how they understood the problem statement and their suggestions on how the solution can be taken to next level with the use of data/algorithms. The second is Technical, which will evaluate innovation, code quality, use of data, scalability of the solution, use of standard libraries, and algorithms used. The final criterion, Results, will measure the accuracy of the solution by comparing it with unseen datasets.
On 15 September, the government will hold a webinar to provide a detailed description of the challenge, participants will learn in-depth, the objective, the process outline, and the expected outcomes of the grand challenge.
From 2017 to 2020, 22,631 accidents were reported due to potholes. These accidents led to 9,223 lives lost and 19,977 injuries. Statistically, 16 accidents occur in India on an average day, and they result in the loss of 6 lives and injure more than 13 people. Road safety is an integral part of road design and the government is investing in tech to reduce human error.
Earlier this year, the Minister of Road Transport and Highways, Nitin Gadkari, announced plans to put AI-based technology in government systems to improve mobility. He identified priority areas for AI applications, including forensic post-crash investigations, the pattern of accidents due to black spots, fatigue indicators, sleep detectors, and advanced vehicle collision avoidance systems.