IIHS Study on Self-Driving Cars and Traffic Fatalities
The Insurance Institute for Highway Safety (IIHS) recently conducted a study in which they envisioned a future where the only cars on the road were of the self-driving variety. The assumption of the study was that self-driving cars would prevent crashes exclusively caused by incapacitated drivers or perception errors since the cameras and sensors on such future vehicles would be flawlessly monitoring the road and identifying potential safety hazards. The technology would be expected to complete these driving tasks better than a human since the machine would be “immune” to incapacitation or distraction.
Crashes Involving Perception Errors and Driver Incapacitation
Crashes caused by driver incapacitation accounted for 10 percent of the total crashes studied, and those that involved perceiving and sensing errors made up 24 percent. Those two types of crashes may have been avoided if every vehicle on the road were autonomous and if all cameras and sensors worked perfectly without malfunction.
That left two-thirds of the remaining crashes as types that still could occur on future roads unless self-driving cars were programmed to assess and avoid other types of errors in performance, decision-making, and predicting.
Self-Driving Cars Have Already Been Involved in Fatal Incidents
One example of the failings of an autonomous vehicle can be found in the Uber test vehicle incident that resulted in a pedestrian fatality in Tempe, Arizona. This incident occurred in 2018 when the driving system was initially unable to correctly identify a pedestrian on the roadside. It eventually did identify the object as a human, but then was unable to predict that the pedestrian would cross directly in front of the test vehicle. The Uber self-driving car did not execute correct evasive maneuvers and ended up striking and killing the pedestrian.
The incident in Tempe was the first known fatality caused by a self-driving vehicle. The IIHS makes the point that the human that was in the test Uber, the monitoring driver, was watching something on his or her mobile phone when the incident occurred. This set of circumstances highlights another potential problem with the autonomous future: When a computer is placed in “control,” many times, humans forfeit their sense of responsibility.
Crashes With Other Causes
Up to 40 percent of other crashes involve driver decision errors, such as active or passive speeding, illegal U-turns, or drifting outside of lanes. The IIHS determined that depending on the circumstances, self-driving technology would not be able to predictably prevent such incidents. This is especially true if companies manufacturing autonomous vehicles continue to place rider preference over safety. In the interest of rider preference, companies may make allowances for stopping, turning, passing, and speeding, which would result in crashes.
The remainder of the crashes studied involved circumstances for which a human or perfect computer system would be unable to predict. Large groups of people may overwhelm a computer system. Adverse weather events can reduce the reliability of the car’s technology. There will also always be times when a tire blows, or a car experiences another spontaneous mechanical failure. Whether a human or a computer is in charge, some events are unpredictable and unavoidable.
Below is a video from IIHS about crash avoidance.
Sacramento Personal Injury Lawyer
I’m Ed Smith, a personal injury lawyer in Sacramento, California. I have been assisting injured Sacramento residents for more than 38 years. If you have been injured by the actions of a negligent driver, reach out to me for free and friendly advice. My local phone number is (916) 921-6400, and I can be reached from out of the Sacramento region at (800) 404-5400.
It is always important to research an attorney before you hire someone. We have many past client reviews and samples of some of our case successes accessible by the below links:
Image : https://pixabay.com/illustrations/car-automobile-3d-self-driving-4343634/
:mm llo [cs 689]