AI to Detect Areas Prone to Waterlogging


Researchers have proposed an AI system to enhance analysis and detection of areas vulnerable to waterlogging

Engineering students from Netaji Subhas Institute of Technology, Delhi have developed an AI system to detect areas prone to waterlogging. This may help metro cities prevent road congestions during monsoon showers.

Initial research

The development of the system is motivated by the damages caused by waterlogging occurrences regularly. The researchers have studied a combination of rainfall, traffic, and location data to predict the gravity of waterlogging issues in the vulnerable areas.

In order to find solutions that are practical as well as feasible, researchers initially studied Manila, the capital city of Philippines, as it has environmental conditions similar to that of Indian cities.

Smartphone-based cab service Uber provided past travel time data that the researchers used to identify areas susceptible to waterlogging. Researchers also considered the elevation data of the area to present a holistic analysis.

Progressive road planning

The intensity of waterlogging was calculated based on the rainfall data and the day of the week, as traffic on weekends and weekdays can vary drastically. The data was run through a neural network to verify the locations prone to waterlogging and even disclose those areas that were unplanned to manage the potential problems.

The data was then fed into an AI system consisting of a neural network capable of deriving patterns from the information fed to it. The system was upskilled by the students to use an algorithm and determine a water logging intensity score to suggest the extremity of possible problems.

Presently, the team plans to modulate the data for Delhi and train the system to reveal patterns on an hourly basis. The system can further be used to not only deliberate strategies for positioning of emergency vehicles but also to understand the effects of festive occasions on traffic. This can eventually lead to developments in road planning.