The main barrier for autonomous driving: we, the people

While autonomous cars in a smart city environment are a great promise for solving urban mobility problems, development of practical applications should take its time and take great care in navigating the dangerous ‘Phase 4’ period, an industry expert warns. In the meantime, municipalities should go ahead and make their mobility plans as ‘closed’ as possible, separating traffic from other parts of public life.

The technology behind autonomous driving, which is a vital part in any future smart city concept, is pretty far underway. Aided by a combination of car mounted and road side sensors, algorithms can make accurate decisions faster than a human can already based on fast analysis of all kinds of data, including the velocity, direction and behavior of other objects on the road.

However, during a presentation in the Polish city in Krakow in the summer, Thomas Walbrun, Director Business Development & Strategy at the Mobility Division of Siemens, warned that full car autonomy will take more time than most would like. Legal challenges have more of an impact than technical ones. But the main problem lies in the required interaction with humans that will remain a staple for years to come. “An in-between phase would be that the driver receives live information and can react to that”, the German said. Still, making the final leap from phase 4 where the AI does most of the driving while a human oversees it, will be difficult. “Everywhere where conventional cars meet autonomous cars we see accidents”, he said. “Every accident is a major setback, so the pace of development is very slow.” He points to the USA where an autonomous car from Uber was involved in a fatal collision in Arizona, postponing any further tests in the state.

“We have to be realistic: it will be in the very far future until we have ‘robocars’ without further human intelligence involved”, Walbrun said.  He added that it is not just safety, and that practicality also comes in play. For example, an artificial intelligence needs to be founded in law, but sometimes extraordinary road situations call for decisions that are not always to the letter of the traffic code. “There will always be a need for human beings in the background who keep an eye on the operations. No city is going to install a christmas tree of sensors every hundred meters <as> it is too expensive due to maintenance and service.”

The situation is different for closed systems though, and one solution would be to separate traffic from all other city activity. “Singapore has by far the clearest plan to solve transportation needs around the world”, he said. “Singapore has a lack of space, even though the city is growing. It already has the highest public transport share for traffic: 67 percent. Compare that to London, where it is 40 percent, or Warsaw, where it is 35.” Singapore’s plan would entail moving all motorized traffic below the city surface, leaving the city space proper for all other activities. That would give much room for autonomous driving solutions. “Singapore has an ambitious target of full autonomous driving cars in all of their townships by 2030.”

 

*) This article was brought to you by Michiel  van Blommestein, editor at Smart City Hub.

3 Comments

  • Well, pedestrians are always gonna be a problem, with the whole automated driving. I don’t see how they would be able to account for people crossing at the wrong time and on the wrong part of the road, or God forbid, people who jump in front of the car to commit suicide. The most hectic part of the automation process would then be, how to deal with security.

Leave a Reply

Your email address will not be published. Required fields are marked *