Each year, thousands of people lose their lives crossing the street. When drivers don’t expect to see pedestrians, sometimes they just don’t see them, even though they’re there. Other times, pedestrians make poor choices, like crossing the road in the dark, which leads to accidents and, sometimes, tragic deaths. Will this reality change once AVs (autonomous vehicles) rule the road?
One of the key selling points for future AVs as well as today’s connected vehicle-safety systems is that they’re safer than human drivers driving without help. However, a new study suggests the industry still has a way to go before this is truly the case. Assuming the study’s findings about today’s autonomous vehicle-safety systems are accurate, what, if anything, might this mean for AV adoption going forward?
AAA released a report that takes a look at automatic emergency braking systems with pedestrian detection. The results are troubling, suggesting automatic emergency braking systems with pedestrian detection perform inconsistently. The research also suggests these systems aren’t effective at night, which, since most pedestrian fatalities happen at night, would mean that this technology is not helping humans behind the wheel as much as they’re expected to.
The AAA study tested and compared the pedestrian detection and automatic braking systems for four vehicles: Chevy Malibu, Honda Accord, Tesla Model 3, and Toyota Camry. In a test that evaluated how vehicles faired at 20 mph when presented with an adult pedestrian target crossing the road, AAA reports each system provided visual notification of an impending collision for five test runs. Two of the four vehicles avoided a collision with the pedestrian target for at least three out of five runs, but the other vehicles hit the pedestrian target every time with very little speed reduction prior to impact.
In increasingly complex situations, such as multiple adult pedestrians crossing the road at the same time or a child pedestrian darting into the road from between two parked vehicles, the survey suggests the performance of these systems failed to impress. In low-light situations, all four evaluated detection systems failed to prevent a collision with a pedestrian. However, AAA acquiesces that each vehicle’s owner’s manual describes system limitations consistent with the limitations discovered in the test.
One take-away from this survey is that these systems are meant to be guides, and drivers must remain engaged behind the wheel and not rely too heavily on driver-assistance technologies. Consumers need to be educated about their vehicle’s safety features, including the limitations of these features.
Another takeaway is a call-to-action for auto manufacturers. These systems need to be better. Going forward, a hurdle to the adoption of AVs is consumer buy-in. Many people don’t trust autonomous systems, and, when comparatively simple technologies like pedestrian detection don’t adequately detect pedestrians, why would they? Whether or not the systems AAA tested truly functioned poorly or if they simply were tested outside of their capabilities, it’s an eye-opener for the industry. AVs will need to be able to detect pedestrians every time and in varied conditions—there is no other option. When will the tech be ready for that next level?
Want to tweet about this article? Use hashtags #M2M #IoT #AV #autonomousvehicles #connectedcar #safety #driverassistance #vehiclesafety #distracteddriving #AI #artificialintelligence #machinelearning #bigdata #digitaltransformation #cybersecurity #blockchain #5G #cloud