BUFFALO, N.Y. – The automotive world is getting amped for robotic drivers. Almost every carmaker has begun testing concept cars that can drive without a human driver actively engaged at the wheel. The technology promises to eliminate human error, the cause of nearly every crash.
But to achieve these goals, driverless cars need to improve in many areas.
The Buffalo car accident lawyers at Cellino & Barnes recently conducted an analysis of motor vehicle incidents and crashes involving automated or partially automated vehicles. They found that computers continue to struggle in several key areas.
- Lighting conditions.
Some vehicles, like the Tesla Model S, have already introduced an ‘autopilot’ mode but the computers aren’t perfect. Tesla’s Autopilot failed to identify “the white side of the tractor trailer against a brightly lit sky.” The error resulted in a deadly crash.
Lighting conditions have always been hazardous for humans too. Roughly half of all crashes occur at night despite there being significantly less traffic on the roads.
Most driverless cars rely on cameras and sensors which often don’t have the same accuracy or depth perception as humans, so nighttime travel has been a challenge for robotic carmakers; but there have been promising safety developments.
Ford unveiled a new device this year that improved autonomous results on the road. Called a LiDAR device, it uses laser beams to construct a 3D image of the environment around a vehicle. Ford recently completed a successful test of the device in complete darkness
- Weather conditions.
While LiDAR can assist computerized cars at night, it struggles to identify the environment in inclement weather conditions. When there’s rain or snow on the pavement, LiDAR struggles to identify important safety markers, signs and even the traffic lanes on the roadway.
“Driverless cars continue to struggle in snowy environments and that should be concerning for anyone who lives in Buffalo,” car accident lawyer Steve Barnes said. “Even though most drivers have experienced these conditions before, snow can be challenging for the most experienced drivers and its unpredictable nature could be a major safety hurdle for carmakers as they develop and improve upon driverless vehicles.”
Once again, Ford successfully tested its automated prototype in the snow. To accomplish this, programmers created 3D maps that allow the car to identify topography or landmarks such as buildings and signs when lane markings aren’t visible.
Black ice poses an even greater problem. Barnes said humans can rarely identify it; and it might be nearly impossible for cameras or sensors to recognize black ice on the roadway.
- City traffic.
One of the most important pieces of technology in any automated vehicle is GPS. But if you’ve ever tried using GPS in Times Square, you could have issues. Tall buildings and other sources of interference can make it difficult for a vehicle to receive a GPS signal and in a concrete jungle like New York people have reported drop-outs.
There are other issues in cities too: there’s increased foot traffic, continual construction, potholes, litter, the list can go on.
With the current technology, automated vehicles can struggle to distinguish the differences between a plastic bag and a puppy.
Bridges seem fairly simple. Almost all of them are straight and predictable. Still, driverless cars are having a rough go.
Uber recently told Bloomberg that its new driverless fleet in Pittsburgh (supervised by humans of course) is having a tough time going over bridges. That’s because the sensors don’t have the ‘hints’ that they’re used to having – things like buildings, signs and objects don’t appear on bridges so it’s harder for the computers to figure out where they are.
Adding to the problem, many suspension bridges are designed to ‘sway’ during earthquakes or high winds. The Buffalo car accident lawyers at Cellino & Barnes said drivers rarely notice the sway but a computer may recognize that its GPS is slightly off.
Human drivers have to know the fundamentals of operating a car: the meanings behind road signs, speed limits, and the consequences. Equally important however is the ability to communicate with other drivers.
Driverless vehicles don’t have that ability yet.
“When two cars approach a stop sign, a driver can wave to the other driver as if to say, ‘you go first,’” Barnes said. “Drivers are constantly communicating with other people around them to avoid potential accidents.”
Earlier this year, Google reported its first accident where the crash was the car’s fault. Safety advocates said it likely could have been avoided with simple human communication.