Whose hands are on the wheel?

Self-Driving CarAssociate Professor Vinayak Dixit makes relationships a priority because he believes that it is through the inter-personal that innovations are created.

One response is to clearly identify AVs to onlookers with improved signage. Just like an L plate or a P plate we need an AV plate. Human behaviour then starts to change because our understanding of the risk associated with an AV becomes part of the interactive relationship and assists in that evolution we need for wide civic application

Associate Professor Vinayak Dixit

In 2013 with funding from the Australia Research Council (ARC), rCITI, CVEN’s transport research hub, created TRACSLab. Already a world leading facility, housing multiple driving simulators networked together, with the ability to connect to traffic microsimulation software, TRACSLab recently formed a field-leading, inter-university relationship with the University of Sydney to investigate the travel behaviour of driverless automated vehicles (AVs).

Funded by insurance industries and transport authorities, Vinayak and his team are asking some fundamental questions raised by AV travel. In AVs, how do people behave? What kind of choices do people make when under risk or uncertainty? How is it to come in and out of autonomous driving? What can a driver do while being in a driverless car? Seated in the TRACSLab driving simulators, drivers in a virtual reality, reveal and record the answers to these questions.

There is still the battle of perception. Incidents involving AVs receive a disproportionate amount of media coverage, which results in a distorted perception of how dangerous they are. “The dangers of AVs are similar to regular vehicles, but often, traffic around an AV on-road become distracted and fearful. AVs are often involved in accidents they have not caused.”

As a committed researcher Vinayak Dixit sees every AV accident as an opportunity for new questions and deeper discovery. “This technology is based on the interaction between the AV and an individual. Currently, an AV finds it difficult to ‘understand’ intent as a human would, because the information exchange between humans is evolved. How could we program this kind of evolved intersubjectivity into an AV?”

Until that time, there are precautions we can take to maximise safety. “One response is to clearly identify AVs to onlookers with improved signage. Just like an L plate or a P plate we need an AV plate. Human behaviour then starts to change because our understanding of the risk associated with an AV becomes part of the interactive relationship and assists in that evolution we need for wide civic application.”

There is also the barrier of price, as AVs are still in the experimental and very expensive stage. One further and very basic challenge to the driverless car is why do we need them? Productivity is the ‘driving’ force for industry and for us city folk sick of wasting our time in traffic. “This technology is about removing human effort, frustration and loss of time around travel, which are all extremely unproductive.”

Then there is the potential for an increase in traffic safety. While this runs at odds to perception, driverless cars will eventually reduce accidents caused by driver fatigue, speeding, inattention and distraction because so much more data is feeding into the car, like the full 360 degree view and the ability to track multiple objects.

Perhaps most rewarding to researchers like Dixit would be the potential of AVs to provide mobility for those who have none or are limited by age or disability. Technology that serves this kind of human freedom is worth pursuing. Dixit knows that driverless cars are a frontier many do not want to cross. But he is a patient researcher. “30 years ago, we accepted auto pilot on planes. Pilots used to make an announcement. Now it is completely accepted as the norm. 85% of flying time is in auto-pilot mode. At our ports, in manufacturing and mining, AVs are common. Perhaps it is just a matter of time for driverless cars.”