Urban India turns to Artificial Intelligence to manage civic woes

AS bharat is poised to be the third-largest world-wide economic system, its cities, too, feature undergone a calm makeover from their colonial aesthetics to a more modern avatar dotted with taller buildings, metro rails, and flyovers. However, one thing that has remained from the bygone era is the occasional appearance of stray cattle in the hustle and bustle of busy streets, and slowing down traffic movement. Worse, they cause deadly accidents.While the problem is ancient, the city administration in Surat has deployed a modern solution of artificial intelligence (AI) to counter this menace.The civic body uses an AI algorithm developed by National Institute of Technology (NIT) Surat to analyse the live video feed from cameras installed at major junctions in the city to detect stray cattle, said Jigan Patel, director of information technology at Surat Municipal Corporation (SMC).“The alert is not issued when cattle are accompanied by a herder. But, whenever stray cattle are spotted, an alert is generated with a timestamp and location,” he said, adding that control room officials inform the on-ground dedicated municipal staff for catching stray cattle, who impound them.Based on the directives of the Gujarat high court, SMC, in 2023, rolled out a free RFID tagging programme containing details of their owners for bovines before implementing a strict anti-stray cattle policy since November that year. The fine amounts increase progressively for repeat offenders, as part of a state law passed in the assembly in 2022 following the same high court directions.So far, 1,545 incidents of stray cattle have been detected, with ₹18.4 lakh recovered as penalties.The same system is used by the city to detect incidents of littering and spitting. SMC commissioner Shalini Agarwal said that this technology-enabled monitoring has helped the city to achieve a high ranking in the central government’s Swachh Survekshan programme.Patel explained that offenders of spitting are often identified using their vehicle numbers detected through the ANPR (automatic number plate recognition) systems. He said so far, 12,120 challans for spitting and 35,382 incidents of garbage throwing were detected through this system, leading to the collection of ₹7.19 lakh and ₹18.4 lakh as penalties, respectively.Just like the age-old problem of stray cattle and encroachments, another issue that plagues Indian cities is that of unaccounted loss of water or non-revenue water (NRW), primarily due to leaks in the water supply network. While globally, NRW stands at 30%, in India, many Indian cities report NRW as high as 50%. The Economic Survey of Delhi 2023-24 found that about 58% of the water produced by the Delhi Jal Board is classified as NRW.However, in Pune, over the last two years, several AI-enabled digital twin solutions have been applied to improve the situation.An official with a major industrial conglomerate tasked with the revamp of the city’s water supply said, as part of a trial, the company’s solution was deployed in the Panchawati zone with a piped water network of around 10km. “By using the system, we could find leakages in our network and illegal users, which have now brought our NRW to 25% currently. We are aiming for further efficiency and bringing it down to 20%,” the official said on anonymity. Based on the success in Panchawati, the system will be deployed in 15 other zones.The digital twin works on data from sensors in the water supply network to create a virtual replication of the water supply network. The AI/ML models process the live data of the sensors, pumping stations, and analyse the household water consumption information to detect leakages and unauthorised usage.The Pune Municipal Corporation (PMC) is also working on expanding an AI-based system to detect unregistered or illegal constructions by using a GIS (Geographic Information Systems)-based system and comparing it with property tax data, an official said.Another city in Maharashtra, Chhatrapati Sambhajinagar (formerly Aurangabad), has developed an AI-based chatbot –– which works in Marathi and English –– for its official website to cater to queries raised by residents.Faiz Ali, project manager at Aurangabad Smart City, said the chatbot, unlike FAQ (Frequently Asked Questions) based bots, is based on various e-governance modules covering the issuance of birth and death certificates, department-related grievance redressal systems, and making property tax and water bill payment systems user-friendly. He said any complaint addressed to the chatbot or the city’s 155304 helpline is automatically sent an SMS acknowledgement stating a timeline. On the backend, the complaint is tagged to the line department official.If the complaint is left unaddressed within two weeks, then the AI-based system flags it for the city commissioner, who reviews the unresolved complaints regularly, Ali said.In Madhya Pradesh’s Jabalpur, the city administration is using a 3D model system of the entire city area spanning 264 sq km, which is then compared with new imagery gathered through drone-based aerial surveys every five years. “Using the system, we can get accurate information on the expansion of buildings. In the latest revision ending March 2024, we found that there were additional structures in more than 20% of the properties,” an official aware of the details said on anonymity, stating this led to an additional property tax collection of ₹30 crore.While these cities have taken a lead in leveraging AI for different use cases, the Centre’s Smart City Mission and Safe City project (through the Nirbhaya fund) has mainstreamed the use of AI-based integrated traffic management systems (ITMS) and surveillance systems. The ITMS systems have AI-based applications to detect red light violations, helmetless usage of two-wheelers or triple riding, wrong-way driving, among others, and identify violators through ANPR. These systems have been incorporated by all tier 1 cities and many tier 2 cities.Similarly, the Safe City project has cameras equipped with facial recognition technology to flag the presence of wanted suspects. New Delhi is soon expected to have an additional 3,500 AI-enabled cameras, along with 200 gunshot detection systems, from October 1 under the same project.While the number of cameras installed increases, their maintenance and usage have come under scrutiny, even during the recent tragic incident involving the death of a Delhi University student, Sneha Debnath. It was widely reported that a series of cameras in and around the Signature Bridge was non-functional.In the week leading up to the incident, HT reported how Delhi’s Public Works Department (PWD) had found many of the existing cameras defunct or in need of urgent repair, while some of the previously acquired cameras were yet to be installed.Debolina Kundu, acting director at National Institute of Urban Affairs (NIUA) –– a think-tank of the Ministry of Housing and Urban Affairs (MoHUA), said AI and ML are vital for predictive analytics that can not only improve service delivery, but also manage disaster risks.In that line, she said under the CITIIS (city investments to innovate, integrate and sustain) 2.0 programme, 21 city-level climate observatories are being set up where real-time pollution and weather data among other datasets will be used along with historical weather data to predict localised weather forecasts and disaster reduction measures. “Integration of different layers of live and historical data is important for this,” she said. Smart city command centres also have a large amount of data that can be further cultivated to gather actionable insights, Kundu added.