What Prince Philip Can Teach Us About Autonomous Cars

Perhaps inspired by Prince Philip and his car accident, a report from across the pond suggests that autonomous cars might be the answer for the declining abilities of older drivers. Rather than getting into a protracted back and forth about how old is too old to drive, can we agree that for every individual, there is a decline in their reactions time with age? Whether that decline is slow or abrupt, at age 70 or 80 is an argument for another day. The report from Newcastle University talks about reaction times under very specific circumstances. By way of a quick disclaimer, the information I am presenting on the Newcastle study comes from a press release; not a peer reported paper; so, take as many grains of salt as necessary. [1]

My disbelief about the autonomy of autonomous cars is detailed here and here. The study looked at response times for drivers at Level 3 autonomy. For a refresher, Level 2 is what you can find in a manufactured car today, lane avoidance, braking, and acceleration assistance. Level 3 is that found on aircraft, the most autonomous transportation systems we deploy. While planes are perfectly capable of flying by themselves, and that includes take-offs and landings, these vehicles require humans to occasional step in, to take control of the aircraft when the unexpected occurs or the situation warrants. In this study, researchers looked at reaction times for drivers who after a few minutes of autonomous driving were distracted, turned away from the steering wheel, feet off the controls, reading from an IPad. 

In clear conditions, it took the older (60 to 81 years old) 8.3 seconds to respond and the younger drivers (20 to 35 years old) 7 seconds. The difference in time resulted in 20 simulated collisions for the older ones and only 12 for the youngin’s. (The press release didn’t provide a denominator)

But does any of this make a difference? Consider the real-world case of the pedestrian killed by the autonomous car in Arizona, with a distracted supervisor in the driver’s seat – just the conditions of the study. He had 2 seconds to react, and by that measure, both the young and old volunteers would have killed the pedestrian. The time that it takes to obtain situational awareness and act upon those findings is just too short for cars. Here is some scientific proof. 

As I have suggested before, if we are not talking about landing or taking off; you have more time to correct the situation in a plane, because you have altitude, there is no similar dimension for cars. 

The typical reaction time is 1.5 seconds when you are paying attention, slightly longer for older drivers, so perhaps a warning system could be helpful. But those situations are going to be limited. The hype remains hype; you are not going to be sitting in the car, reading or watching TV, you are going to have to be actively supervising at least for the critical periods – getting to and from the highways. In the meantime. if older people have slower reaction times they should increase the distance between cars - at 60 miles per hour that would be 8 car lengths just to account for the differences in reaction times documented in this study. 

Source: Are self-driving cars really the answer for older drivers?

[1] Subsequent to writing the article I did receive the peer-reviewed paper. It does not materially change any of the information provided or my conclusions. Here is the direct source: Investigation of older drivers’ take-over control performance in highly automated vehicles in adverse weather conditions.IET Intelligent Transport Systems. DOI:  10.1049/iet-its.2018.0104  The visualizations of avoidance tell the entire story.