An in-house first-of-its-kind innovation by two officers has helped the Indian Army overcome the age-old problem of tracing blind tank and motor rounds that fail to go off during firing but have proved catastrophic taking lives of many civilians when it goes off accidentally.

 Lt Colonel VK Pandey and Major Akshat Chaudhary’s prototype ‘UXO Detection System’ has been successful as it traced 102 artillery and infantry shells washed away from an Eastern Command Depo in Sikkim during Teesta flash floods last October.

“The Artificial Intelligence (AI)-based software, fed with over 2,250 to 3,000 images per ammunition from different angles, is linked to an unmanned aerial platform (UAV) which flies and scans an area of operation to detect images of Unexploded Ordnance (UXO) or rounds that did not fire, the Army is looking for. The software model analyses video feed from the drone to detect the presence of UXO,” Lt Col VK Pandey told businessline.   

It will detect, alert the user, and plot the location of detected UXOs on Google map, thereby creating a digital footprint of the area. This data can then be utilised by troops for timely destruction of the blind rounds preventing chances of fatal accidents, he explained. The model is flexible and can detect even scattered ammunition protruding 20 percent above the ground since the drone is trained on an enormous data bank to pick similar images from a distance in air, the Lt Col said, sharing insights into the capacity of the system.

Undergoes test

The system has been tested at practice firing ranges to check its efficacy. One in say 1,000 rounds due to technical error may not explode even after firing and lies scattered on the ground which is dangerous and can get triggered if fiddled with, Major Chaudhary said. 

However, this system is not meant to identify small arm bullets. 

As of now, the job of collecting and destroying UXO is being done manually in the Army, which is labour intensive and time-consuming, and nearly impossible for the jawans to sweep an area of firing normally spread in about 10 kilometres completely without leaving unattended any blind ammunition, the Major elaborated.

Now the same thing can be done by one quadcopter, its operator and a laptop and its user, Lt Col Pandey stressed. Comparatively, the time taken is also very less.

To scan one km by one km area would require about six soldiers and more than eight hours and the same sweeping can be done in one hour by the UXO Detection System, Pandey narrated. The cost is not huge either since the UAVs are already available in units and high-resolution cameras as payloads on quadcopters are not very expensive too, he added.

Pandey and Chaudhary had a eureka moment in their lives when the unit they were posted in Jharkhand thought of finding a solution after a blind round explosion in Ranchi last year consumed two lives, they said.

The techies in olive green uniforms had the right academic background to brainstorm for an innovative output --   while Pandey is Master’s in AI and is from the corps of engineers, Chaudhary holds a BTech degree in Computer Science and belongs to corps of signals. 

 The concept started in March, 2023, post Ranchi ammunition explosion and the data integration on the software took one to two months before they started conduction initial trials. They faced many challenges especially during the operation carried out from October to December 2023 to locate ammunitions lost due to Tista floods, the two officers observed. It took them one month to locate scattered blind rounds of artillery and infantry given the area was random, Pandey recalled.

This model can be later tweaked to identify personnel and men trapped or lost to avalanche, they said when asked if the system is capable of handling such human miseries as well.

Related Topics
comment COMMENT NOW