Chunar stn first to get AI software for rly track safety

Allahabad: Months after the unmanned crossing along Chunar Chopan section of Allahabad division of North Central Railway (NCR) became the last to be eliminated in the nation, Chunar railway station became the first in the country where the ‘Hanging Part Detection System’ (HDPS) was installed.
Comprising an extremely sensitive camera with an artificial intelligence-based software, the system detects any object found to be hanging 155 millimetre above the railway track and avert accidents like derailment.

The system is so designed that it can detect any object hanging from a wagon running at a speed of 130 km/hour, the maximum permissible speed for trains running in Allahabad division including India’s fastest train Vande Bharat Express.

Public relation officer of NCR’s Allahabad division, SK Gupta, said, “Among several features, this system alerts officials when a hanging part is detected. Any object detected above up to 155mm of the railway track will be treated as a hanging object except the wheel.”

The HDPS system was installed along the yard at Chunar railway station in November, 2018 by senior divisional mechanical engineer Ajit Kumar Singh and his team. However, the authorities started recording the system’s performance only from March 15. Gupta said that on March 18, the system had sent out alerts although it was later found that the hanging object was actually a plastic bottle.

“If the system detects any object hanging it sends out alerts to a number of officials through SMS. After receiving the SMS, the authorities can check the report generated by HDPS for any unusual readings. These readings also include data like temperature of the wheel and axle (as it is integrated with existing hot box detector), speed of the train and number of wagons,” said Gupta.

The hot box detection system records the temperature of the wheel and the axle of the train.

Moreover, the HPDS collects images of the tracks, performs an imaging analysis and sets a continuous communication with controller of hot box detection system. After receiving the alert, NCR officials would be able to access the recorded video feed of the railway track.