I think DUI/DWI kill way more people than autonomous vehicles.
As a software engineer I know all about risks of using software. However, those in a modern (non-hands free) car are relatively small and nowhere as dangerous as using autonmous driving features. I hate that stopping on the slightest of inclines causes the âHill startâ feature to switch on; I wish I could switch it off as more than times than not is causes the engine to stall when I try to drive off.
No because I doubt that it would be safe under UK law and most likely to fail its annual MOT test.
Nobody is advocating here for autonomous vehicles; we are talking about driver assist systems where the human remains fully engaged and responsible for safety.
Can you tell me what scenario you fear with comma.ai that is risky?
A key benefit to me for comma.ai is that the stock cruise control for my vehicle turns on (or off) both lanekeeping and speed management simultaneously. Often I want to slow down on curves or approaching an intersection. comma.ai lets me turn on lanekeeping alone while I manage the speed myself. I find that to be very useful for local driving and tend to turn on full cruise (lanekeeping plus speed management) only on major highways.
I find this ability to turn on lanekeeping only to be a very desirable feature. I cannot do it with my stock car but can do it with the comma.ai. Do you think this is dangerous? If so - why?
Your âHill startâ example is a good one. I suspect you could create a fork of OpenPilot to disable the Hill Start feature of your car if there are ECU commands via the CAM bus which turn it on and off; would you use it if someone else wrote it, or would you consider writing it yourself if you have the time?
1 Like
Lanekeeping is always on in my vehicle because I pay attention to the road and the markings thereon. As to cruise control there is only one reason for my using it â on those sections of motorways with a speed reduction in force and average speed cameras to check compliance. It all comes down to what is the offence of âdriving without due care and attentionâ.
Is that a serious statement or satire?
Are you suggesting there is no value to lanekeeping as a feature in a car?
1 Like
I guess that I am. For several reasons including what are the insurance implications of using âdriver aidsâ like this especially when these are after-market hacks to the vehicleâs electronics? And this recent report on the UKâs Royal Navy fiddling with wiring does nothing to install confidence.
Another reason is I subscribe to the statement of the late Enid Mumford, professor at the Manchester Business School, back in the 1970s or 80s on computerisation âjust because you can does not mean you shouldâ.
Ultimately it is all about âdriving with due care and attentionâ.
I am confused about what the Royal Navy story has to do with the safety of lanekeeping features.
Both Consumer Reports and IIHS are well-regarded sources regarding auto safety - they have released several studies similar to this one which strongly support lanekeeping and related advanced driver assist features.
You are of course welcome to your opinion - but it seems to be based on anecdote or conjecture and seems to ignore data.
Or one could look at the research into Swedenâs mandatory use of side lights on vehicles, which started off saving lives but driver complacency crept in from this passive driver aid such that since the early 2000s the death rate from RTAs increased plus some of the original data was invalid.
https://www.sciencedirect.com/science/article/abs/pii/000145759500015R
Apples and oranges
If you have similar data on Advanced Driving Assist Systems, then indeed we should look at that data