AI-based iRASTE helps reduce road accidents
The pilot programme began in July 2022 with 190 AI-based advanced driver assistance system (ADAS) devices and 10 driver monitoring system (DMS) units installed on 200 buses
Hyderabad : A locally designed artificial intelligence-based system has reduced road accidents involving RTC buses by 40 per cent, as compared to the buses that did not have the system, its developers have stated.
The system uses a road-monitoring camera paired with an alert mechanism placed in front of the driver. While the camera monitors road conditions and provides data, the alert system warns drivers three seconds ahead of potential incidents.
The pilot programme began in July 2022 with 190 AI-based advanced driver assistance system (ADAS) devices and 10 driver monitoring system (DMS) units installed on 200 buses.
The study took place between April 2023 and March 2024, covering three major interurban highways, from Hyderabad to Kodad, Pullur and Adilabad.
“RTC crash statistics for April 2023 to March 2024 show that buses equipped with ADAS reported 40 per cent fewer accidents compared to non-ADAS buses on the same routes,” said Govind Krishnan, programme manager, iRASTE (Intelligence Solutions for Road Safety through Technology and Engineering).
iRASTE is an AI-based advanced driver assistance system (ADAS), developed by IHub-Data, a Central government-funded technology innovation hub located at IIIT Hyderabad, in collaboration with Intel and INAI, the Applied AI Research Centre of IIIT Hyderabad.
“Our AI systems are trained to adapt to different climatic and time conditions such as fog, mist and night driving. We’re continuously feeding live road data to improve their accuracy,” iRASTE’s Krishnan said.
Real-time alerts notify drivers of unsafe behaviours such as lane departures or following too closely, helping them improve their driving habits.
CEO of IHub-Data Dr Jay Mookherje shared the results of the pilot project with Deccan Chronicle. “We have trained more than 500 TGSRTC drivers on using this system. During the project, the buses equipped with the ADAS units reported 40 per cent fewer accidents than the non-ADAS buses from the same depots on the same routes. With India leading the world in road fatalities,” he stated.
Using live data from the ADAS equipment from the buses, IHub-Data is working to improve its accuracy. “Our AI systems are trained to adapt to different climatic and time conditions such as fog, mist and night driving. We’re continuously feeding live road data to improve their accuracy so that the drivers could get increasingly accurate alerts,” said Krishnan.
Vouching for the results, TGSRTC Miyapur depot manager Ramaiah Singham said their drivers were quite satisfied with the ADAS and DMS units. “Drivers face different types of stress while driving. Several times, they were saved by the alerts from iRASTE equipment. The number of accidents has been reduced considerably thanks to this equipment and our drivers are happy,” he explained.
“The iRASTE Telangana project demonstrates the potential of AI-based systems in improving road safety. If widely adopted for public transportation, our ADAS could significantly reduce road accidents and fatalities. It is important that we remove the tag of the country with the most accident fatalities,” added Dr Mookherje.
Drivers receive four types of alerts
The Intelligence Solutions for Road Safety through Technology and Engineering (iRASTE), developed by IHub-Data, is an AI-based advanced driver assistance system (ADAS)
Forward collision warning (FCW) to prevent hitting vehicles in the front
Geadway monitoring and warning (HMW) for maintaining a safe distance
Pedestrian collision warning (PCW) if pedestrians, cyclists or two-wheeler riders come in front of the vehicle
Lane departure warning (LDW) for changing lanes without turning on the indicator or drifting away from roads.
The driver monitoring system (DMS), another vital feature of iRASTE, uses facial recognition technology to track eye movements, head position and facial expressions, detecting signs of drowsiness or distraction.