five reasons why autonomous cars aren’t coming anytime soon - laser sensor

PITTSBURGH—
In the world of driverless cars, Pittsburgh and Silicon Valley are the bustling centers of development and testing.
But ask those who are involved.
When we can see vehicles carrying passengers in every city, you get an almost universal answer: not very fast.
10 years is an optimistic assessment.
Many others say researchers have tried to overcome some obstacles for decades.
The vehicle itself will be in a limited, good debut
The map area inside the city extends outward.
In Arizona, a fatal accident in an Uber self-driving car has caused the accident to progress slowly, mainly because it has damaged public perception of vehicle safety.
The company slowed down its research to be more careful.
Google's Waymo, for example, decided not to launch a fully autonomous journey --
The car-hailing service in the Phoenix area, at least for the time being, will rely on human backup drivers to transport passengers.
Here are the problems that researchers must overcome when they start riding without human driving: snow and weather are big enough to cover the road surface, snow blocks the line of sight of the lane used by the vehicle's camera to find the road.
So far, researchers have not found a solution to this problem.
That's why most of the tests are done in a warm environment.
Climate like Arizona and California.
Heavy snow, heavy rain, heavy fog and sandstorms can hinder the camera's sight.
The beams from the laser sensors can bounce back from snowflakes and think of them as obstacles.
The radar can see through the weather, but it can't show the shape of the object the computer calculates what it needs.
"It's like losing a part of your vision," said Raj Rajkumar, professor of electrical and computer engineering at Carnegie Mellon University . ".
Greg McGuire, director of the MCity autonomous vehicle testing laboratory at the University of Michigan, said researchers are looking at using laser sensors with different beam wavelengths to penetrate snowflakes.
Software is also being developed so that vehicles can distinguish between real obstacles and snowflakes, rain, fog and other conditions.
However, many companies are still trying to complete the difficult task of driving on sunny days with stable traction.
"Once we are able to run reliably in these systems, then we will start working to expand to those more challenging conditions," said Noah Qi, head of system security at Uber. driving cars.
On a global scale, road markings are different and may not even exist.
Lane lines are not standardized, so vehicles in each city have to learn how to drive in a different way.
Sometimes there is no curb to help the vehicle judge the width of the lane.
For example, in the industrial "Strip area" of Pittsburgh, many
The driving vehicle was tested and the city drew a line in the narrow lane to indicate where the vehicle should be parked.
Sometimes, these lines are very far away, buildings are very close to the street, and if the self-driving car stops on the line, they do not see the traffic at the intersection.
A workaround is to program the vehicle to stop the line and crawl forward.
"Is it better to do double stations?
Asked Peter Rand, president of Argo AI, a self-driving car company that Ford has invested heavily in.
"It is not easy because the intersections are different.
"For years, when dealing with human driving problems, self-driving cars have to deal with people who don't always follow the rules. They double-
Park or walk in front of the car.
Recently in Pittsburgh, when his car stopped at the right turn, an Argo backup driver had to take over, and when it could not immediately decide whether to go around, he blocked an intersection.
Parked delivery truck
"Even if the car may eventually find some problems, it is also a shared space that is socially unacceptable," Rander said . ".
Humans also have eye contact with other drivers to make sure they are moving in the right direction, which is also the direction of self-driving cars.
Plus some people are hostile to robots.
It is reported that people have been harassing Waymo's automatic test car near Phoenix.
In December, the Republic of Arizona reported that police had recorded at least 21 cases in the past two years, including a man who opened fire on a Waymo van, as well as people who cut tires and threw stones.
A Jeep forced the van on the road six times.
In the face of oncoming traffic, without a green arrow, turning left is one of the more difficult tasks for human drivers and one of the reasons for many accidents.
There is the same problem with self-driving cars.
Waymo CEO John clafchick said in a recent interview that his company's vehicles are still experiencing occasional problems at intersections.
"I think we are facing challenges as well," he said . ".
"So sometimes unprotected lofts are extremely challenging for humans and sometimes for us.
"Consumers accept last year's deadly Uber crash near Phoenix, doing more than pressing the test pause button.
It also upset consumers who will one day be asked to ride their own bikes.
Driving a vehicle.
An investigation after Uber's crash showed drivers were reluctant to give up control of computers.
A AAA survey found that 73 American drivers were too scared to fully self-drive. Driving a vehicle.
This ratio is from 63 at the end of 2017.
The self-driving car company is displaying information on the screen of the Test passenger about the direction of the vehicle and what the sensor sees.
Krafitch of Waymo said that the more people they have, the more they trust the vehicles.
"When they become more confident, they rarely look at the screen and they are on the phone, relaxing or sleeping," he said . ".