NEW DELHI: The Indian army is betting big on artificial intelligence (AI) to deal with threats. It has deployed several units of AI-powered intelligent surveillance systems to the northern and western borders of the country. This reduced the need for manual monitoring.
In addition, the Indian Army has deployed an AI-based Suspicious Vehicle Recognition System in eight locations in the Northern and Southern theaters.
An AI lab has been established at the Military College of Telecommunications Engineering where AI projects undergo extensive internal testing before being handed off to a production agency for deployment.
“We have an AI surveillance system used at multiple locations where inputs from thermal images, cameras and electro-optical devices are integrated and we are able to give real-time inputs to field formations,” said sources. This provides soldiers in the field with real time.
“It has features like motion sensor alarms, so when a device detects motion, an alarm appears on the TV screen,” sources said.
The military has also deployed an AI-based suspicious vehicle tracking system in insurgency-affected areas where vehicle registration numbers are noticed and matched with Department of Surface Transportation data. “It gives real-time input to soldiers in that particular area to identify if the vehicle registration number is correct or incorrect,” sources said.
The Indian Army has also developed an AI-based cervical cancer detection system which has been deployed in some hospitals.
“We made our algorithms in-house in our institutions where we established a center of excellence for artificial intelligence. We formed a pool of specialists, who started working with the field formations trying to have assets that are deployable in the field. These can be surveillance-based systems, human resource management systems or logistics systems. All are based on algorithms that we have developed in-house,” sources said. The advantage of the in-house research center, sources said, was that “we used to get the data from the formations in the field.”
“We used to create the algorithms internally and then feed them back to training. Again, it was an ongoing effort where we interacted with the formations on the ground. We have 145 asset deployments that have been created on this basis”, said sources.