Combating Urban flooding with Internet of Things and Artificial Intelligence

It’s no denying that we have been witnessing unprecedented leaps in natural disasters across the world, let it be wildfires, COVID pandemic, floods, drought, extreme weather, and so forth. Every country, developed or developing, takes note of it and be ready for the next cycle of occurrence. However, Pakistan is an exception and it is evident by the current situation of urban floods and power outages. It’s a long debate as to why precautions are not being taken or why it lacks coping strategies when it’s apparent that the floods pose a problem every year. This article deviates towards the short/long term solutions for the current situation of urban floods with the help of Artificial Intelligence. I’ll deal with the most pressing concern, i.e. urban floods due to heavy rainfall.

Heavy rainfall has caused urban flooding due to which cities like Karachi and Hyderabad have become either a swimming pool or a sewer depending on the area you live in. The worst part is that weather advisory has issued a warning for more thunderstorms while we are experiencing a situation that can best be described by a scene in the Jumanji movie. Well, let’s talk about solutions. There are two approaches one could follow to solve the current situation. The first is to use an already developed system “CENTAUR” acronym for Cost-Effective Neural Technique to Alleviate Urban Flood” which uses artificial intelligence techniques to manage the flow of water in sewers. However, it requires the installation of gates in the sewer to control the flow. It’s a kind of a load balancing mechanism such that the gate can hold the water in sewers for one area if it is not flooded already to its maximum extent and divert it to other parts of the city. Technically, the method uses remote sensors and Fuzzy logic to control the gates. The pilot installation of the CENTAUR system has shown promising results in Toulouse, France, and Coimbra, Portugal. Although it is considered to be a low-cost solution still expecting our governments to spend thousands of dollars is like waiting for a miracle to happen. Moreover, this solution is good for small floods but during larger floods, the solution may tremble a bit. Concerning the current situation, the containment strategies cannot be implemented straight away (we have way surpassed the time for damage control), however, the solution for the next year would be to use smart manholes as adapted by the Japanese and Chinese governments. To be honest, it is a quite rough sketch of the system in realization but it gives a certain understanding of how the system would be.

The depicted system can monitor the water levels in sewers on a real-time basis. Furthermore, it can also help to monitor the flow of the water, leakage in pipes, or any other anomalies. The system with smart manholes can be implemented with existing infrastructure with a very low-cost in comparison to the former solution. The sensor deployment and its communication will be the main issue in the manholes but the benefit of doing so is that the administration will have all the information needed at their disposal along with some predictions regarding the rise of water-levels in sewers, rainfall, and alert system, accordingly. With the help of AI, massive accumulated data collected from hundreds if not thousands of smart manholes will make it possible to rapidly forecast the water-levels, the areas where sewerage systems need repairing instantly, effective utilization of equipment, and the anomalies in underground wiring and cables. The system initially would take some costs for sure but afterwards, it can reduce the costs by efficiently improving the sewer maintenance system. Finally, in terms of the current situation, the sensors can sense if a cold-blooded animal is floating around in the sewers.

In summary, the former solution would provide a kind of long-term solution but it needs infrastructure to be installed before the system could be put at work and the later solution can be implemented straight away with sensors, communication equipment, and battery installations. In both cases, the AI component is crucial as both the system heavily relies on the prediction system and data analytics based on the acquired data. Nevertheless, it will be the government/administration at the end who have to take necessary actions but at least they would be notified two to three hours earlier before the mishap happens. We need to learn how to incorporate and leverage AI techniques to improve the quality of life and at the same time improve the ecosystem. I hope this article could be considered as a cornerstone for adopting technology to solve our societal problems.