Remember meForgot password?
    Log in with Twitter

article imageSelf-driving vehicles need long development before general use

By Ken Hanly     Feb 5, 2019 in Technology
While areas such as Pittsburgh and Silicon Valley are actively researching the development of autonomous vehicles, they are not likely to be seen actually carrying passengers or be released for sale for years to come.
How long will it take for release of autonomous vehicles?
Even an optimistic assessment puts general use of the vehicles at ten years. Some even think it will be decades. There are numerous obstacles facing the development of the vehicles.
Vehicles already operate in limited well laid out areas within cities. Their use will spread out from these areas. The fatal accident of an Uber autonomous vehicle in Arizona last March damaged public perception of the safety of the vehicles. As a result research became more careful. Although in the Uber crash that car was in self-driving mode there was also a driver in the car as well. The accident may have been in part caused by the actions of the cyclist victim.
Google's Waymo, as an example of safety measures taken after the crash, decided it would not launch a fully autonomous ride-hailing service in the Phoenix area. For now it is using human backup drivers in the service. Here are some of the problems researchers face.
Snow and weather
When snow becomes deep enough to cover the lane markings that the vehicle's cameras rely on to find their way, the vehicle cannot determine where it is in a lane very well. So far researchers have found no solution to this problem. This is why most research is being done in places with a warm climate such as Arizona or California.
Heavy snow, rain, fog and even sandstorms block the view of cameras that the vehicle use for sensing where it is. Light beams sent out by laser sensors can bounce off snowflakes mistaking them for obstacles. Although radar can penetrate the weather it cannot perceive shapes necessary for computers to decide what an object is.
Software is being developed that can differentiate real obstacles from heavy rain, snowflakes, fog and other conditions.
Pavement lines and curbs
From place to place markings on roadways vary and on many roads do not even exist. Even when there are lane lines they are not standardized. Vehicles must learn how to drive differently in various cities, countries, and conditions.
In addition to these problems, on many roads there are no curbs by which vehicles can judge lane width.
Dealing with human drivers
Human don't always obey the rules of the road. Autonomous vehicles are often not equipped to deal with the situation. Humans often double part for example. In Pittsburgh recently, an Argo self-driving vehicle stopped during a right hand turn and was blocking an intersection as the AI could not decide whether to go around a double-parked delivery truck. Fortunately the vehicle had a backup driver.
Some humans do not like autonomous vehicles. In Phoenix people have been reported harassing Waymo's autonomous test vehicles. Police in suburban Chandler have documented at least 21 instances in the past two years, and one includes a man who waived a gun at a Waymo van, and others in which tires were slashed and rocks thrown at vehicles. A jeep forced vans off the road six times.
Left turns
In places where there is no green arrow, it is often problematic for autonomous vehicles to decide if it is safe to turn left. Even humans find this a bit of a challenge. Waymo CEO John Krafcik said: "I think the things that humans have challenges with, we're challenged with as well."
Public acceptance of autonomous vehicles
Many people fear riding in autonomous vehicles and many others fear being on the road with them or worry as pedestrians or cyclists. Recently AAA found that fully 73 percent of Americans are too fearful to drive in an autonomous vehicle. This is up from just 63 percent in late 2017.
However Krafcik from Waymo pointed out that the more people ride the more they trust the vehicles. Passengers in the test vehicles are shown information on screens about where vehicles are headed and what their sensors are seeing. Krafcik said over time people looked less at the screens and were on their phones or even sleeping.
On website Quora, plant engineer Justin Culmo notes: "Price: will be the biggest inhibitor of adoption. Self-driving will only be in higher-end cars for a long time, and better versions (levels of autonomy) will come out, first, in the more expensive cars."
Culmo predicts that driving enthusiasts will stay away from self-driving cars. Many people will still treasure driving around in their older cars and sportscars. No doubt many will trust driving their own cars over riding in an autonomous vehicle.
More about autonomous vehicles, selfdriving electric vehicles, autonomous vehicle research
Latest News
Top News