On a collision course with driverless cars

Automated Vehicles (AVs), also known as driverless cars, will bring with them a raft of changes to the way we move around cities.

Some pundits predict that because AVs will be programmed to protect humans, people will deliberately walk and cycle in front of cars, exercising their dominion over these new, subservient vehicles. If so, automobile travel would be hampered, potentially allowing walking and cycling to reclaim the public space gradually lost to the motor vehicle over the past century.

While this vision (a utopia to some, a nightmare to others) sounds plausible, there is little evidence in the history of the motor vehicle to suggest that AVs will make cities more friendly for walking and cycling (although perhaps fewer people will die). AVs may simply repeat the history of the motor vehicle, with public space once again cleared to make way for people travelling in cars at the expense of walkers and cyclists.

Some studies anticipate that AVs will make better use of road networks than human-driven vehicles by travelling closer together, making more efficient route choices and by communicating with each other.¹ This suggests that traffic congestion will disappear. However, other studies predict the number of vehicle kilometres travelled is likely to increase substanitally,² suggesting we may end up with roads saturated with even more motor vehicles.

Traffic simulations predict AVs will interleave seamlessly at intersections without the need to stop for traffic lights; however, these simulations conveniently ignore the rights of people to walk and cycle in this mechanised future. Technology companies have proposed that people will use mobile apps that operate as beacons to alert AVs of our presence. But this approach envisages a world in which everything that moves, including children, dogs and native wildlife, will need to emit a beacon to stay safe.

For insight into the likely impact of AVs, let's examine the first pedestrian fatality involving an AV, which occured in March 2018. The preliminary report from the United States National Transport Safety Board found that, despite the vehicle detecting the pedestrian about six seconds before the crash, the vehicle did not adjust to avoid the collision. 

Some reports suggest the software had been “tuned” to be less sensitive to objects on the road so that ride quality was better for passengers. This prioritisation of occupant comfort over the safety of walkers and cyclists is concerning if unsurprising.

While Uber’s automated vehicles currently utilise a human operator to monitor performance and intervene if necessary, the system was not designed to alert the operator when intervention is required. The assumption that a mainly passive vehicle operator can be relied upon to intervene when necessary has always been questionable, given the likelihood of distraction. In this case, it turned out to be fatal, with reports indicating the Uber operator was watching TV before the crash.

If roads are dominated by computer-controlled vehicles, the only form of transport controlled by humans will be walking and cycling. This means road safety messaging and police enforcement may focus more heavily on walking and cycling as they struggle to remain relevant in an automated world. Given cycling in Australia and the US has already been cast as a brave and somewhat foolhardy endeavour, a greater focus on the dangers of walking and cycling by safety agencies is likely to further marginalise these healthy forms of transport.

We have two paths ahead. Down one path, we cede more public space to the automobile and fence off roadways to prevent pedestrians from ruining the highly-efficient streams of automated vehicles (an approach somewhat akin to the introduction of jaywalking laws 100 years ago). Down the other path, we reprioritise public space to favour walking and cycling, thereby improving the liveability of our cities and the health of its inhabitants.

• Tony Arnold is a researcher at the Institute of Transport and Logistics Studies

The choice is up to us.

1. Fagnant D, Kockelman K (2015), Preparing a nation for autonomous vehicles: opportunities, barriers and policy recommendations
2. Truong LT, De Gruyter C, Currie G, Delbosc A (2017), Estimating the trip generation impacts of autonomous vehicles on car travel in Victoria, Australia