AI-based alerts help prevent elephant collisions against trains
As forest cover decreases and human-animal interactions increase, especially on railway tracks, railway and forest officials are turning to surveillance and alert systems based on artificial intelligence as a solution.
K RajendranSep 09, 2024Wayanad and Palakkad, Kerala; Coimbatore, Tamil Nadu
As we expand our road and rail networks, especially through forest areas, it spells problems for the wild animals. Several villages on the fringes of the Western Ghats, such as Kottekkad village in Palakkad district of Kerala and Madukkarai village in Coimbatore district of Tamil Nadu, face this problem. As these villages lie close to the forest, wild elephants often wander into the villages in search of food and water. This year alone two elephants have died after collisions with speeding trains.
As earlier efforts, such as underpasses, have not yielded good results, the state governments are trying artificial intelligence to avert elephant collisions with trains.
Tackling man-animal interactions near railway tracks
Southern Railway has already constructed two underpasses at Railway Line B at Madukkarai forest range in Tamil Nadu to prevent elephants from getting killed on the railway tracks.
“Even after constructing underpasses, herds of elephants are crossing railway tracks and we are very scared,” said L Jayalakshmi of Puthupathy village who lives just 1 km away from the elephant underpass.
Indigenous people like the Paniyas and Malaval Makkal living beside the railway tracks in Puthupathy village have also been facing many challenges.
While those with enough financial means protect their houses with concrete walls and electric fences to ward off wild elephants, the indigenous people living in huts are badly affected as the tuskers just destroy the huts easily.
“The strange thing is that authorities are more concerned about keeping elephants safe from the speeding trains. They don’t seem to think of the livelihood of the poor,” said Shobhi, a herder in Puthupathy village.
Shobhi has been grazing her goats on the sides of the railway tracks where plenty of shrubs and bushes grow. She feels safer near the tracks as the elephants go into the villages in search of food and water.
AI-based alerts prevent collisions
The Tamil Nadu forest department recently started an artificial intelligence-linked elephant surveillance and warning system along the tracks in Madukkarai village. Cameras with AI have been installed over 12 towers along the railway track.
The cameras provide live images of wild elephants to the control room that monitors the feeds round the clock. Simultaneously, the AI-based alerts are passed on to the railway officer concerned, and thereafter to the locomotive pilots who can slow down the trains to prevent collision with elephants.
But some of the locomotive pilots are sceptical about the AI-based surveillance system.
“Even if we get a proper alert about elephants crossing railway lines, we cannot apply the brakes beyond a certain level. Hence we can’t solely rely on this technology,” said a locomotive pilot on condition of anonymity.
Since October 1, 2023, an AI-based elephant surveillance system has been providing alerts. The system covers a 7 km stretch of railway line between Valayar in Kerala to Ettimalai in Tamil Nadu.
“Over the last six months, we have provided 718 AI-based alerts. As a result, not a single collision has been reported in this stretch of railway line in the six months,” said K Priyadarshini, surveillance officer at the monitoring cell located at Solakarai village in the Madukkarai Forest Division.
Kerala experiments with AI-based fences
Now the Kerala forest department has decided to experiment with the country’s first AI-based smart fence project, called Ele-fence. It has been initiated in Chelakkolly village in Wayanad district in Kerala, as three human beings were trampled to death by wild elephants in the recent past.
The Ele-fence system includes live monitoring of wild elephants, their protection and prevention against collisions. The fence connected with AI-based trip wires would provide early warnings of elephant movements. It also incorporates a lasher belt technology, which would prevent elephants from physically breaching the fence.
“Through this system we will ensure warning as well as protection. Ele-fence has been developed after deeply delving into elephant behaviour in typical conflict-prone situations,” said Parrakkal Mohanan Menon, CEO of Kochi-based White Elephant Technologies that has installed the fence.
Is technology enough to prevent man-animal interactions?
Now Southern Railway has sanctioned the implementation of a cutting-edge elephant intrusion detection system along the Kottekkad-Madukkarai route, spanning Kerala and Tamil Nadu with an allocated budget of Rs 15.42 crore.
“We are committed to utilising advanced technology to safeguard wildlife, while also ensuring seamless operation of train services along the elephant corridor,” said Arun Kumar Chaturvedi, divisional railway manager, Palakkad.
Wayanad district, which extends over 2,125 sq km, is famous for its lush green hills, tropical climate, flora and fauna. About 20 percent of the district comprises tribal population of 14 different indigenous communities, especially Kurichia, Panica, Adiya and Kattunaikkas.
Tribal people have invariably depended on forests for livelihood and they are the most affected by man-animal interactions. The tribes are not sure if technology would prevent the conflicts.
“Changes in the forest landscape compel wild animals to venture into human habitation. How could this core issue be resolved with Ele-fence?” asked Girijan Gopi, principal scientist at the M.S Swaminathan Research Foundation, Wayanad.
According to the Comptroller and Auditor General report tabled in Kerala Assembly on July 11, from 1,811.35 sq km in 1950, forest land in Wayanad district has reduced to 863.86 sq km in 2021.
Hence, as many experts and affected people point out, without stopping the loss of forest cover, AI-based systems might provide some relief from man-animal interactions, but technology in itself cannot be a permanent solution.
The lead image shows a surveillance system at Madukkarai, which sends AI-based alerts to the railway authorities and locomotive pilots. (Photo by K Rajendran)
K Rajendran is a journalist based at Thiruvananthapuram.
This article has been written with the support of Mumbai Press Club-N R Prakash Fellowship for Science & Technology for 2024.