Engineering safe behaviour to save lives

So said School of Civil Engineering PhD candidate Stephanie Ballon-Romero, when asked why she chose to study transport engineering.

Because thoughts of death were never too far away when she was growing up in Peru – with inadequate and developing road infrastructure, many fatal accidents happened.

“I felt that this was something I could improve, and so I decided to study civil engineering at the Universidad Católica Santa Maria,” she said.

“Most people think of civil engineering as just designing structures such as bridges, but it can also include analysing transport systems and how people travel between places – whether by foot, bike or vehicle.

“I enrolled in every transport course I could in my undergraduate degree and enjoyed it. In fact, my honours thesis was The study, simulation and optimisation of traffic flow along Jorge Chavez Avenue in Arequipa.

With such a passion for her subject, it was only natural that Ballon-Romero would pursue further studies. Having learnt Chinese as part of her program – second languages are highly regarded in South America – she successfully applied for a government scholarship to complete a Master of Engineering in China. Under the supervision of noted traffic analyst Professor Hongmei Zhou at the Dalian University of Technology, she then completed the paper, An extension of the theory of planned behaviour to predict pedestrians’ violating crossing behaviour using structural equation modelling.

“The ‘theory of planned behaviour’ tries to understand why people do certain things, especially reckless behaviour like smoking or crossing a road when the light is red,” Ballon-Romero said.

“It’s a psychological approach widely used in the development of behavioural campaigns to reduce fatalities and improve safety. I applied this theory to understand why pedestrians behave recklessly and which factors influence their behaviour, such as how other people behave, how people think others should behave, how easy or difficult is, and cultural ‘norms’.

“I found that people in China generally try to do the right thing so as not to ‘lose face’ – there is much peer pressure to conform – whereas in more ‘individualistic’ societies like Peru and Australia, this is not always the case.”

The ‘structural equation modelling’ aspect of her paper was the mathematical model she developed to analyse the survey she conducted on 260 people in Dalian.

When creating her survey, Ballon-Romero used questions such as ‘Do you feel safe crossing the road at a red light?’ and found that if people thought jaywalking was safe, they would be more likely to do it.

“You need to design the right questions to be able to measure results and to be then able to use the tool widely,” she said.

A similar model was validated in Australia, and she found that about 30 per cent of pedestrians behave recklessly.

Ballon-Romero’s research is of vital interest to government agencies and private companies such as Tesla, Waymo and iMOVE Australia looking towards the future of transport as vehicles become more and more automated.

“Just think of current features such as cruise control and satellite-based navigation (Sat Nav) that we all take for granted: this is the first level on the scale for automated vehicles (AVs),” she said.

“Level 2 may include more complex interactions such as the vehicle automatically carrying out steering, acceleration and braking.

“Levels 3 (conditional automation) and 4 (high automation) will require less and less intervention from drivers, until full automation at Level 5, where no driver is required behind the wheel at all. This reality is closer than we think – for example, Singapore already has some areas of the country dedicated to AVs only.”

Which means that understanding the interaction between pedestrians and vehicles fully reliant on algorithms and machine learning is a big issue.

“We can adapt though – in the Netherlands, for example, where there are lots of bike riders, traffic lights at intersections are being removed altogether, and people rely on negotiation only.”

And the good thing about AVs is that, like humans, the more they drive, the better they get at driving. They learn more about traffic conditions, cycle lanes and pedestrian/bike rider behaviour because they add so many millions more photographs to their system from the many cameras on their top, sides and back, which are constantly scanning the environment.

“AVs will actually be safer than human-controlled vehicles and will bring immense benefits in terms of work, play, safety and eased traffic congestion,” Ballon-Romero said.

In the meantime, we need to find out more about how people currently behave on the road to make the experience better and safer for all because, at present, the world has yet to fully embrace the concept of AVs. Attitudes will need to change if we are to enjoy their benefits.

“I think the introduction of AVs will be a gradual process and will be linked to the launch of the more environmentally sustainable hybrid and electric vehicles,” said Ballon-Romero.

The next step for Ballon-Romero is to conduct virtual reality tests (using VR headsets) to expose people to AVs in a safe environment and to observe how they, as pedestrians, react. The test results could then be used to recommend to AV manufacturers where best to place cameras – perhaps on bumpers or windscreens instead of hoods.

Alternatively, she is considering a detailed study of bike riders and measuring how they interact with AVs on cycleways.

“Do we need to teach behaviour or modify transport to feed actual behaviour? That is the question I’d like to answer,” said Ballon-Romero.

And once she does that, she’ll know that her research has made a difference.