The Karnataka State Road Transport Corporation (KSRTC) plans to use Artificial Intelligence (AI)-based technologies to limit road accidents and improve passenger safety in buses. According to a report, the corporation recently floated a tender for the implementation of an AI-powered Collision Warning System (CWS) and Driver Drowsiness System (DDS) for 1,044 buses. CWS will provide features like forward-looking collision warnings (FLCW), lane departure warnings (LDW), and virtual bumper. It will also generate real-time alerts. This is probably for the first time in the country a state-run bus corporation is using technology on a large scale to reduce accidents. Other state-run bus corporations are also waiting to adopt this system.
The tender is likely to be finalised by the end of June 2021, the report said. KSRTC officials said the FLCW system will identify an impending collision and inform the driver that they have entered an unsafe distance zone. An official noted that this would help the driver prepare to take the necessary action to avoid a collision. The system will provide real-time alerts to warn the driver against impending collisions. AI-based camera sensors will provide the detection of a vehicle from a sufficient range of at least 150m at any speed so that it can effectively warn the driver.
When minimum safe distance is not maintained, an alert will be generated. This minimum safe distance is based on a calculation of the time-to-collision (TTC) with the vehicle ahead including 2/3 wheelers, pedestrians, and cyclists. The officials added that the alarm will be initiated at a TTC of up to 2.5 to 3 seconds, be operational at a vehicle speed range of up to at least 120kmph, and generate both visual and audible alarms. It will also notify the driver when lane marks are not available.
DDS will check its drivers from dozing off at the wheel. It will monitor the driver’s eye movements and sound a warning alarm in case they appear sleepy. AI-based CCTVs will watch the facial behaviour of the driver. It will also alert the KSRTC central control room if the driver ignored the alert. This will be helpful for night services, said an official.
In April, OpenGov Asia reported that the Indian Institute of Technology, Ropar (IIT-Ropar) had developed an algorithm for driver drowsiness detection using machine learning and computer vision. The researchers said they used computer vision algorithms to extract facial features such as eye closure and yawning as well as machine learning techniques to effectively detect driver’s alertness. It is an industrial and academic challenge to develop drowsiness detection technologies.
Multiple techniques have been developed in recent years. One method is where the driver’s operation and vehicle behaviour can be monitored by the steering wheel movement, accelerator or brake patterns, vehicle speed, lateral acceleration, and lateral displacement. Another set of techniques focuses on monitoring the physiological characteristics of the driver such as heart rate, pulse rate, and electroencephalography. The third set is based on computer vision systems, which can recognise the facial changes occurring during drowsiness.
The first method is limited by the type and model of the car. The second method though with more accurate results has widely been downplayed due to the impracticality in deploying it on a large scale, as well as its intrusive nature. The third method is a very promising one, which the researchers have followed and developed a model on the same.