Research shows autonomous cars hitting stopped objects, not staying in lanes
August 10, 2018
The self-driving vehicle movement has reached another roadblock, one that apparently some autonomous vehicles cannot see. In a report released by the Insurance Institute for Highway Safety, some Level 2 autonomous vehicles failed to stop for stationary objects, failed to stay in lanes, or experienced safety issues in other ways.
IIHS evaluated the 2017 BMW 5-series with Driving Assistant Plus, 2017 Mercedes-Benz E-Class with Drive Pilot, 2018 Volvo S90 with Pilot Assist, 2018 Tesla Model 3 and 2016 Model S with Autopilot. Road and track tests studied the effectiveness of adaptive cruise control and active lane-keeping features.
In studying adaptive cruise control, vehicles were subject to four series of track tests to examine how they respond to another vehicle in front of them stopped and exiting lanes. One test had the cars going 31 mph toward a stationary vehicle target with cruise control off and autobrake on. Both Teslas hit the stationary target.
With adaptive cruise control on, the BMW, Mercedes and both Tesla vehicles came to a slow, gradual stop, with Tesla cars braking earlier. However, the Volvo S90 braked just 1.1 seconds before impact to avoid collision, resulting in a forceful stop.
Another test had the cars following a lead vehicle that changed lanes to reveal a stationary inflatable target ahead with about 4.3 seconds to impact. With cruise control activated, none of the vehicles struck the target. The Volvo still had a more forceful brake than the other test cars.
However, results were less favorable for the technologies away from the track and out on the road. Every vehicle except the Tesla Model 3 failed to respond to stopped vehicles ahead on an actual road.
One researcher was driving the Mercedes E-Class at 55 mph on U.S. 33 near Ruckersville, Va., with both active cruise control and lane-keeping features activated. With no vehicle in front of her, the E-Class detected a pickup truck stopped at a traffic light ahead. Unfortunately, the detection system lost sight of the truck and continued at its current speed. The researcher had to hit the brakes herself to avoid a crash.
“At IIHS we are coached to intervene without warning, but other drivers might not be as vigilant,” the researcher said in the report. “(Adaptive cruise control) systems require drivers to pay attention to what the vehicle is doing at all times and be ready to brake manually.”
Other vehicles hit the brakes too much. Tesla’s Model 3 slowed down 12 times in 180 miles. Researchers noticed that seven of those times coincided with tree shadows on the road. Other times involved oncoming traffic (in the correct lanes) and vehicles crossing the road far ahead.
Looking into lane-keeping features, test vehicles were subjected to six trials with three different sections of road. The Tesla Model 3 was the only vehicle to stay in the lane in all 18 trials. The Model S overcorrected on only one trial.
Both the E-Class and S90 stayed in their lane in nine of 17 runs. BMW’s 5-series technology stayed in its lane in only three of 16 runs. In some cases, the test vehicle would follow a lead vehicle switching lanes, rather than staying in its original lane.
IIHS concluded that the outlook is promising for potential safety benefits of adaptive cruise control. The institute was less impressed with active lane-keeping, noting the evidence for safety benefits is not as pronounced as cruise control technology.
For autonomous vehicles as a whole, IIHS said fully self-driving vehicles are far away from becoming a reality.
“We’re not ready to say yet which company has the safest implementation of Level 2 driver assistance, but it’s important to note that none of these vehicles is capable of driving safely on its own,” said David Zuby, IIHS chief research officer. “A production autonomous vehicle that can go anywhere, anytime isn’t available at your local car dealer and won’t be for quite some time. We aren’t there yet.”