The South African National Roads Agency Limited (Sanral) is investigating how machine learning (ML) can be harnessed to improve road safety, reduce congestion and inform infrastructure development.
Sanral notes that its innovation arm, the Technical Innovation Hub (TIH), is leading this project.
Founded by Sanral in 2019, TIH is described as a think tank, which aims to be at the forefront of harnessing technologies to inform, improve and accelerate road safety on the South African road network.
In a statement, Ruan Van Breda, mechatronics engineer at TIH, explains that ML can be used to detect and segment objects in a camera image – every frame in a video is analyzed as a still image.
“These objects can then be classified based on pre-trained image classifiers. In the road environment, it allows to detect and classify different types of vehicles, pedestrians, different types of animals, cyclists, etc. ”
Van Breda believes the possibilities are endless, adding that a lot of data is already available for these types of classification.
These genres can be further extended through the creation of custom data sets and training classifiers, to be able to distinguish, for example, between slow traffic and a road accident, he says. It can also be used to create new classification classes based on unique experiences or the requirements of the road authority; for example, fire or protest detection, foreign objects such as rocks and tire detection.
The information can then be used to activate the appropriate response through the road incident management system, remedy the situation and inform road users – in real time, he says.
“You can also watch how these different objects interact with each other; for example, to detect unusual vehicle behavior, such as a vehicle stopping on the highway. In addition, information can be deduced about the interaction between several elements such as cars and pedestrians.
“If a vehicle moves on the side of the road and stops and pedestrians move towards the vehicle and enter the vehicle, it can be considered informal pick-up. As more and more data is collected, these trends can either assist road authorities in planning for infrastructure such as drop-off / pickup points or help law enforcement to stop illegal pick-ups if considered. as a security risk.
According to Sanral, even though technology of this nature presents “significant” risks, every effort is made to understand how to use the technology effectively while maintaining strict compliance with the law regarding the privacy of road users. .
Some of the ways to mitigate these potential privacy risks are to use strict security and access controls. In addition, the data can be anonymous at the point of capture.
“Although this technology is still in the exploratory phase in South Africa, it already has languages in countries like China, where they are using machine learning to incorporate facial recognition for law enforcement. They are able to identify the driver of a vehicle and instantly issue fines, if that driver does not have a valid driver’s license.
“Fines can also be issued automatically for people walking around or entering restricted areas. As with any technological breakthrough, there are pros and cons, and in a complex society like South Africa, for now, let’s watch and learn, ”concludes Van Breda.
Sanral plans to use machine learning to improve road safety
Source link Sanral plans to use machine learning to improve road safety