07-21-2022, 01:04 PM
(07-21-2022, 12:48 PM)jamincan Wrote: Part of the problem is that anything involving human factors can have very subtle effects that make it less obvious that an intervention improves safety or increases risk. Lots of cars have additional safety features now, such as lane-assist, blind-spot detection, and so forth. On the face of it, these should make vehicle operations safer, but if the driver learns to rely on those features, can they be relied on to be a full substitute for an attentive driver? Will the blind-spot sensor detect the cyclist over-taking on the right in the bike lane?
This is a good point.
Quote:This goes into factors like street-design as well, where a safer street according to engineering guidelines is almost always very unsafe when you actually examine statistics. The discrepancy is largely due to human factors - people will drive the speed they feel is safe, not the posted the limit, as an obvious example.
In that sense, "asking people to do better" is precisely the solution we need. Or more precisely, forcing people to do better is the solution we need. Make it so they have no choice but to slow down and be attentive, and they will slow down and be attentive.
When I say that humans can be adapted, I don't mean "asking people to do better", I mean specifically using interventions against people that actually change behaviour.
The reason that environment changes are often preferred is because a) they are politically easier -- shocking I know and b) they scale better.
But things like ASE are a form of human behaviour modification that is effective. It solves some of the scale issue with automation, but it's obviously politically difficult.