HYDERABAD: A team at International Institute of Information Technology Hyderabad has developed a smart wearable safety system designed to detect workplace accidents in real time and alert supervisors within seconds, addressing delays in emergency response in high-risk industrial environments.
The system, developed at the institute's Centre for VLSI and Embedded Systems Technology under the leadership of Abhishek Srivastava, targets sectors such as thermal power plants, oil refineries and construction sites, where workers operate in dispersed and hazardous conditions.
Real-time alerts to bridge reporting gaps
Industrial accidents often go unreported immediately, especially in large facilities where workers are spread across wide areas. The wearable device seeks to close this gap by automatically flagging incidents without relying on manual reporting.
IIT-H opens NABL-accredited lab for drone propulsion testing
The device, designed as a compact belt-mounted unit, is connected to a base station and a central monitoring system. It tracks worker location, detects hazardous gases, monitors vital signs, and identifies falls.
"If an accident occurs, the device sends an alert within seconds. The central monitoring station then triggers an audible alarm, enabling timely medical response," Srivastava said.
The system is built around the concept of the "golden hour" the first 60 minutes after a serious injury, when prompt care can significantly improve survival chances. The innovation has been termed the GoldAid system.
Health monitoring integrated into workflow
Beyond accident detection, the device continuously monitors key health indicators such as heart rate, body temperature, oxygen saturation and blood pressure.
Workers can log baseline health data before starting their shift. Any deviation during work triggers alerts, allowing supervisors to intervene early.
"This marks a shift from reactive safety to preventive care embedded within daily operations," Srivastava added.
Machine learning enables fall classification
The system uses accelerometers and gyroscopes combined with machine learning models to analyse movement patterns. It distinguishes between routine slips and critical falls, including estimating fall height.
According to the team, the models have been trained on diverse real-world activities and achieve over 98% accuracy. Processing is done directly on the device, eliminating reliance on cloud systems and ensuring instant alerts.
Designed for large-scale industrial deployment
To enable deployment across large facilities, the system uses LoRa (Long Range) wireless communication, allowing stable data transmission with low power consumption.
The wearable has undergone field testing at a thermal power plant in Ramagundam and multiple construction sites in Hyderabad, where workers used it during routine operations.
Backed by government, nearing commercial rollout
The project is supported by the Department of Science and Technology and has progressed to an advanced development stage, with multiple research publications and patents.
The team is now working on refining the system for large-scale industrial adoption.
SOS feature enables manual emergency alerts
In addition to automated detection, the device includes an SOS button that allows workers to send immediate alerts if conscious and in distress.
The monitoring system continuously displays worker data, ensuring that supervisors can respond quickly to emergencies.
The developers said the system aims to reduce dependence on manual reporting and ensure that accidents are detected and addressed without delay.