Tamil Nadu launches AI-based system to avert elephant deaths near railway track in Coimbatore

As per a statement issued by the state government, the Coimbatore forest division experienced a significant increase in Human-Elephant Conflict (HEC) in recent times.

It was noted that the rising elephant population, disturbances in migration corridors, increased developmental activities in and along the animal migratory paths, changes in land use patterns and agricultural practices, and anthropogenic pressure led to increasing instances of HEC in the region.

The elephants in the Coimbatore forest division are seasonal migrants, who move from Nilgiris and Sathyamangalam to forest areas in Kerala.

“Specific locations, such as Walayar, Bolampatty, Anaikatti reserve forest, Gopinari reserve forest, Hulical, Jaccanari slopes, Nilgiris Eastern Slope reserve forest, Soolakkarai, Singapathy, and Iruttu Pallam blocks, are some of their (elephants’) preferred retreats during the rainy season,” an official said.

Here is a short film on Madukkarai Project in Tamil Nadu. The AI enabled Project aims to prevent death of elephants on railways tracks in the area. Kudos to the TN Forest team of the Coimbatore Division for working 24*7 and executing the task flawlessly #MadukkaraiElephantProject pic.twitter.com/ojsmyQQsOg

— Supriya Sahu IAS (@supriyasahuias) February 9, 2024

It was noted that in the Coimbatore forest division, the elephants had strayed out about 9,000 times between 2021 and 2023.

“One of the major conflict issues in the Coimbatore division was elephants crossing the railway track in Madukkarai range and train hit accidents. There are two railway tracks passing through Soolakarai beat and Bolampatti Block-I Reserved Forests in Madukkarai Range. This forest area shares its boundary with Kerala forests along the Walayar River and is regularly visited by elephants. Unfortunately, 11 elephants have died so far, due to collisions with trains since 2008, including young calves and juveniles,” the statement reads.

It was noted that despite the best efforts of the officials, such incidents could not be drastically reduced and so an AI-enabled autonomous surveillance system to track the elephants and avoid accidents was proposed to be implemented.

After conducting a field investigation, a seven-km-long railway track was identified as the ‘most vulnerable’ and the government allocated Rs 7.24 crore for the installation of the artificial intelligence system.

How can AI reduce human-elephant conflicts?

As per officials, the AI-based surveillance system has 12 high towers fitted with both thermal and normal cameras, installed at strategic locations along the track in Bolampatti Block-I forest. The towers are installed 500 m from each other, covering all the important elephant crossing areas, giving 150 m coverage on either side of the track for early detection of animal movement.

The sensed data is automatically transferred to the control room that processes the data on a real-time basis. The front-line staff of the forest department along with the technical team will be available at the control room on a shift basis to monitor animal movement and inform loco pilots through calls, SMSes, and alerts.

In addition to that, hooters and digital display alerts are also placed near the track for loco pilots.

“The forest department and the railway officials work in sync to prevent accidents based on information generated by Artificial Intelligence. This Artificial Intelligence-generated data not only prevents accidents, but it also gives valuable data on elephant movement, elephant behaviour, profiling of individual elephants, and further morphological and behavioural studies for future decision making,” the statement adds.

Additional Chief Secretary Supriya Sahu took to X to share a short film on the project in the state.

She noted that the alert mechanism has been divided into ‘three zones’ – Red, Orange, and Yellow – based on the projected distance between the track and elephants with ‘Red’ being closest to the track. She added that each zone will have a different system of alerts.