Are self-driving cars ready for the road?

170412 driverless full
Inside the RobotCar autonomous vehicle.
John Cairns / Oxford University Images / Science Photo Library

The majority of Australians under 50 are ready for driverless cars, according to a recent poll, but safety concerns delay self-driving vehicles from hitting Australian roads.

Many manufacturers are hard at work developing their own variations on driverless vehicles. Ford has committed to the task, Tesla has bounced back from an early driver fatality, and Uber already has a fleet of self-driving cars operating in Pittsburgh in the US.

The navigation systems generally combine map-driven GPS data with radar, laser and camera-based technology to pinpoint the exact location of the vehicle at any given moment. 

In theory, the vehicles also use their camera systems and laser scanning to identify objects and people, read traffic signals and signs, detect road infrastructure and scan large areas of their surroundings. 

In Australia, it’s not technology holding us back from seeing self-driving cars on the road. According to Jonathan Roberts, a professor in robotics at the Queensland University of Technology, integrating these vehicles into society is the tricky part.

“It is very hard to create a reliable self-driving car,” says Roberts. “The main risk at the moment is that we will have roads with both human-driven and self-driving cars. The mix is very hard to deal with.” 

Roberts says the associated risks could be mitigated by allocating designated self-driving car areas, but there are still some issues with the vehicles in development. “At the moment, the systems being tested on our roads cannot cope with bad weather, for example.”

Roberts says the first vehicles to hit Australian markets are much more likely to be driverless, automated pods, similar to Google’s prototype, which ferries people back and forth in a restricted area near the tech giant’s headquarters in California.

This pod technology is more like public transport – operating in controlled areas, with predetermined pick-up and drop-off zones – and could have wide applications for delivery and transport across industries from food to medical equipment. 

Experts find it useful to distinguish between five levels of automation in self-driving and driverless cars. Level 1 automation is already in widespread use: cruise control. Level 2 refers to systems that allow a driver to hand over pedals and steering to the vehicle but take control back at any time. These systems already exist in some high-end cars. 

 “Level 3 means a driver need not be alert in some situations,” Roberts explains, such as in well-known environments or on freeways, but must be prepared to take over if the unexpected happens. Some companies consider this undesirable, he says, and are aiming to move straight to Level 4.

“Level 4 means that the driver can completely check out in certain driving conditions and there is no expectation that they will be able to regain control.”

The Tesla-based vehicle in development by Bosch and the Victorian Government is more in the vicinity of level 3 automation, designed to navigate roads with and without human input. It features technology to assess the alertness of the driver, to combat accidents caused by fatigue.

Roberts says it’s difficult to predict when the technology will hit our streets. But when it does, the uptake will depend on how well it interacts with the current infrastructure on our roads.

“One of the key concepts of using driverless cars is that the roads should not need to be enhanced in any way,” Roberts explains. “The business case would likely fall apart if large changes to our roads were required.”

Please login to favourite this article.