Autonomous cars won't fix traffic problems

A geospatial mapping firm and an automotive industry analytics consultancy have joined forces to publish a white paper outlining the complexity of merging autonomous cars with driven vehicles on the world's road networks as population inevitably grows.

In essence, the white paper – published by HERE and SBD Automotive – argues that for autonomous motoring to be a panacea for our clogged traffic arteries, the world's governments and car manufacturers must subscribe to 'collaborative' autonomy... cars that communicate with other cars and the broader environment.

And beyond that, two issues faced by autonomous motoring proponents will affect how little or how much self-driving cars can reduce road trauma and/or congestion. One is the rate of advancement in future autonomous-car technology; the other is the willingness of drivers to be early adopters.

A recent fatality in a Tesla has highlighted the first issue – that semi-autonomous cars won't be sufficiently sophisticated to negotiate the world's roads now... and likely not in the future either, as population density increases.

“Autonomous cars have the potential in the long-term to revolutionise mobility and radically improve the safety of our roads,” said co-author of the study Andrew Hart, Director at SBD.

“However, the journey towards the fully autonomous car is full of potholes, which may create short-term pains in unexpected ways. The automotive industry and road authorities will need to work carefully together to navigate around these potholes, in order to gain the trust of consumers and reap the societal benefits of this new technology.”

The authors of the white paper insist that the best strategy for easing congestion with autonomous motoring is for car companies, governments and other stakeholders to work closely together. At the present, there's a shrapnel approach to the technology – and if a car can't communicate with other vehicles around it or the infrastructure (traffic lights and level crossings, for instance), then it will be hindered from making its way efficiently through traffic.

Consider, as one example, two cars approaching each other on a suburban side street with a car parked at the point where the two moving vehicles will meet. If the two cars can communicate with each other, one can insist on right of way, on the basis it doesn't have to move to the wrong side of the road. But if the two cars can't communicate with each other, they will both slow more than necessary until somehow the impasse is resolved.

In that sort of circumstance, autonomous cars may prove to be slower and less efficient than two drivers.

As we look forward to introducing this new technology populations continue to grow rapidly. According to the local arm of HERE, within 15 years New South Wales is expected to accommodate another two million residents, and Victoria's population will double from its current level.

“Managing the impact of autonomous cars on traffic congestion will not be smooth sailing, particularly in the face of mixed fleets on road networks," says Brent Stafford, Director of HERE APAC.

“Bridging the difference between driverless and traditional vehicles is vital and requires a shift in mindset from each-to-their-own to collaborative autonomous vehicles. This shift needs to occur globally at industry, government and societal level.

“We are already seeing collaboration occur with the Sensoris data standard that will enable driverless connected vehicles to prepare for changing conditions and hazards well before the vehicle, be it a truck or car, can see them. A problem remains however where cities remain trapped in data silos. Traffic signal controller data must be able to ‘talk’ to cars not just talk to the traffic management centre, and buses must know that the train is running three minutes late. The data needs to be made widely available for network optimisation.

“Imagine a future where a pothole is detected by your vehicles’ sensor and automatically notifies the road authority, in turn notifying the agency of the need for repair. Where a stalled car automatically alerts the relevant repair agency, while directing other vehicles to alternate routes. These are the realities that scaled collaboration amongst all stakeholders can enable.”

Stafford expects the period between 2021 and 2036 will be the danger period. During that period there'll be a changing mix of new-tech and traditional cars on the roads, and if it's hard for an autonomous car to judge how a human driver will behave in traffic, the reverse is at least equally true.

After 2036, Stafford believes autonomous cars will be cleverer, the infrastructure will be better networked with the traffic, and there'll be far fewer human drivers piloting vehicles on the roads.

The white paper is available to download for further information.