"To add to the already extensive research done on aircraft pilot"
Studying human factors has done more to reduce aircraft crash rates since the early 1970s than all other changes combined. It's one of the reasons why most airlines no longer directly recruit ex-military pilots (wrong mindset for safe bus driving - a military pilot will push on regardless in marginal conditions instead of diverting. There might be a 90% success rate when doing so, but that 10% is a bitch - and there are far too many cases of "hero pilot abilities saved the day" when "hero pilot's" lack of judgement and fixation on target was directly responsible for putting the flight in danger in the first place.)
Google is absolutely right to be studying human factors and the NTSB's crash studies always find a chain of events. Human factors start right back at the level of road design/layout and ensuring that markings are done sensibly.
The problem is that the vast majority of "road engineers" have no actual training in road safety other than the mechanical stuff and a huge number of "no brainer" assumptions(*) about what's safe or not safe result in decisions being made which actually make roads _less_ safe.
In particular there's a major tendency to look at XYZ "problem" on a road and not realise that it's a symptom of a larger problem. You see this in towns, where panic about people being able to cross the road safely results in extra pedestrian crossings, lights, parking restrictions and fencing being installed, which perversely have the opposite result to what's intended(**). The REAL problem is usually "why is there so much traffic on _this_ road, why is it travelling so fast and why aren't drivers paying attention to conditions?" and the proper answers are usually "Find a better route for that traffic/encourage use of XYZ existing better route and/or slow it down dramatically along with making sure drivers are paying attention to surroundings instead of focussing on the traffic light ahead"
(*) Assumption is the mother of all fuck ups.
(**) to wit:
Crossings usually result in slightly higher levels of car vs pedestrian incidents and they cluster around the crossings.
Fences, parking restrictions, lights, etc _ALL_ result in traffic going faster and drivers getting tunnel vision (ie, paying LESS attention to what's on the footpath, etc) - and fences are particularly lethal if a cyclist/motorcyclist/pedestrian is pinned between it and a car - which usually results in victim blaming by people who think that fences are there for safety (they're not, they're there to guide pedestrians to a "safer" crossing point and if they're being jumped or bypassed, or a documented hazard to other road users they must be removed immediately - almost all of the 2000-2016 "HGV vs bike" deaths in London were a direct result of cyclists being crushed against fencing as an example, vs being able to fall/escape onto the footpath when the HGV turned across their path.)