-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Phantom braking (TACC/AP) [Not AEB]
- Thread starter cciebryan
- Start date
for the seat heating icon use your index and middle fingers to anchor at the bottom of the screen and then use thumb to hit the screen icon
I tend to anchor fingers on left of screen and use my thumb to navigate most of the other functions and that seems to work well for me. Stabbing the screen with a floating finger is a recipe for frustration
pdk42
Active Member
Did a trip of 320 miles today, mostly motorways. I used AP a lot and had about 6 incidences of phantom braking, always when passing a lorry. The visualisation clearly shows the lorry hopping to the right into the same lane as the car when it’s in reality remaining in its own lane.
Potential software fix to avoid rear end collisions: flashers automatically on when passing lorries?
Assholes that tailgate might not have strong reading skills (as well as other intellectual challenges)
sadly true. I must admit that if anyone is behind within a few car lengths I drop out of autopilot 'just in case'....
Wol747
Active Member
for the seat heating icon use your index and middle fingers to anchor at the bottom of the screen and then use thumb to hit the screen icon
I tend to anchor fingers on left of screen and use my thumb to navigate most of the other functions and that seems to work well for me. Stabbing the screen with a floating finger is a recipe for frustration
We rest our case.....
Glan gluaisne
Active Member
Part of the problem is the length of time your eyes need to be refocussed on to the screen, both to ensure your finger touches the right place on the screen and to ensure that the screen has correctly responded to that touch. The time that your eyes are off the road is a lot longer than is the case for a physical control, where tactile feedback from your fingertips provides assurance that the control has operated as intended.
In cockpit HMI design this was something learned decades ago, with the feel of controls being made different so that many could be both found and operated by touch alone. One helicopter procurement I managed included large LCD displays both for normal flight instruments and a range of sensor and weapons data presentations and actions. Touch screens were initially suggested, but the challenges in safely operating these that were found by trained crew during simulation sessions ruled them out. Driving is potentially more hazardous than flying, in terms of the need to stay eyes-out. Much of the time when flying, a several second sweep around the instrument panel presents little or no hazard. Just a couple of seconds eyes-in when driving could well be lethal.
In my view, car driving control design needs to be undertaken to a higher standard in terms of the HMI than for flight controls in an aircraft, because of the risk posed by spending too much time eyes-in. Also, the effect of the design on the performance of those that wear glasses needs to be more carefully considered. Many people need glasses for short sight correction, and as they age will inevitably also need correction for age-related long sight.
Varifocal glasses are a very good solution, that work well for instruments and controls that are in the central line of sight, but they work far less well for peripheral vision. The clear focus zone with the eyes looking at an angle to the side and down is much smaller than it is when looking ahead and down. The result of this is that focus accommodation time increases, it takes longer for the eyes to resolve a clear image when looking sideways through varifocals than it does when looking straight ahead. This adds to the time that the driver has to be eyes-in to operate a touch screen control that's to the side, adding to the problem
Before this comment attracts yet another insulting response, I should add that my corrected vision is absolutely fine for both driving and flying, and that I still fly regularly and maintain my licence. In fact the last type I flew, shortly before the lockdown tiers put a stop to it, had elliptical wings and 1600hp of supercharged V12 power provided by a Rolls Royce Merlin 66, so not the sort of thing someone with dodgy eyesight, or poor hand-eye coordination might fly.
In cockpit HMI design this was something learned decades ago, with the feel of controls being made different so that many could be both found and operated by touch alone. One helicopter procurement I managed included large LCD displays both for normal flight instruments and a range of sensor and weapons data presentations and actions. Touch screens were initially suggested, but the challenges in safely operating these that were found by trained crew during simulation sessions ruled them out. Driving is potentially more hazardous than flying, in terms of the need to stay eyes-out. Much of the time when flying, a several second sweep around the instrument panel presents little or no hazard. Just a couple of seconds eyes-in when driving could well be lethal.
In my view, car driving control design needs to be undertaken to a higher standard in terms of the HMI than for flight controls in an aircraft, because of the risk posed by spending too much time eyes-in. Also, the effect of the design on the performance of those that wear glasses needs to be more carefully considered. Many people need glasses for short sight correction, and as they age will inevitably also need correction for age-related long sight.
Varifocal glasses are a very good solution, that work well for instruments and controls that are in the central line of sight, but they work far less well for peripheral vision. The clear focus zone with the eyes looking at an angle to the side and down is much smaller than it is when looking ahead and down. The result of this is that focus accommodation time increases, it takes longer for the eyes to resolve a clear image when looking sideways through varifocals than it does when looking straight ahead. This adds to the time that the driver has to be eyes-in to operate a touch screen control that's to the side, adding to the problem
Before this comment attracts yet another insulting response, I should add that my corrected vision is absolutely fine for both driving and flying, and that I still fly regularly and maintain my licence. In fact the last type I flew, shortly before the lockdown tiers put a stop to it, had elliptical wings and 1600hp of supercharged V12 power provided by a Rolls Royce Merlin 66, so not the sort of thing someone with dodgy eyesight, or poor hand-eye coordination might fly.
Wol747
Active Member
I guess your M3 is called "Spit"?
I could not agree more with your comments. I have long said that the level of accurate computation involved in FSD is at least an order of magnitude more than even for autoland. The same apples to the controls - IF they ever get to full FSD in the level 5 sense - and I have doubts about that - then touchscreen controls can make sense. But the combination of the touchscreen system now with manual or monitored autonomy is potentially lethal.
The gentleman who "instructed" you how to use a screen control entirely misses the point and in any case I would guess that he has a model 3, where I imagine you can use the side of the screen as a rest - in the S you can't. In any case, your points are not invalidated.
I could not agree more with your comments. I have long said that the level of accurate computation involved in FSD is at least an order of magnitude more than even for autoland. The same apples to the controls - IF they ever get to full FSD in the level 5 sense - and I have doubts about that - then touchscreen controls can make sense. But the combination of the touchscreen system now with manual or monitored autonomy is potentially lethal.
The gentleman who "instructed" you how to use a screen control entirely misses the point and in any case I would guess that he has a model 3, where I imagine you can use the side of the screen as a rest - in the S you can't. In any case, your points are not invalidated.
I get what you’re saying. From a safety point of view I’m guessing we don’t have enough data yet to determine whether the touch screen in Tesla’s are causing more accidents. Real world results can be contrary to what might seem obvious (like the cycle helmet debate). Sure, it takes a little more attention to navigate the screen but most drivers are likely aware of this and take due care.
If you really want to focus on safety then visibility obstruction caused by car A-plillars should be an obvious redesign consideration. As should the law about taking hands free calls. While we’re at it we should probably ban having passengers in the car as this would no doubt be a distraction for the driver and cars should definitely be limited to 70mph
ah yes model 3. My bad didn’t realise the situation was much different in the S. it must be due a refresh soon hopefully the design will evolve
Glan gluaisne
Active Member
It will take years for anything to show in accident statistics, if it ever does. Loss of driving concentration/distraction due to poor control/display ergonomics isn't data that's routinely captured, unless there's a serious systemic problem detected. Lots of cars have poor control layouts, it's not unique to Tesla at all.
One issue is that motoring journalists, and customers, are often deceived by the initial appearance, and don't begin to appreciate the downsides of some design aspects until after a fairly long period of use. It's very easy to refuse to be critical of things, on the basis that you love the design and styling and that any problems are just a matter of getting used to doing something differently.
After over a year of ownership, I'm now well over the stage where any poor control usability could be put down to a lack of familiarity - using that touch screen with my left hand, whilst wearing glasses, just isn't something that's either quick or easy. The recent changes that have made the speed display far less visible, as well as effectively hiding all the light icons behind the driver's left hand, haven't improved things, either.
One issue is that motoring journalists, and customers, are often deceived by the initial appearance, and don't begin to appreciate the downsides of some design aspects until after a fairly long period of use. It's very easy to refuse to be critical of things, on the basis that you love the design and styling and that any problems are just a matter of getting used to doing something differently.
After over a year of ownership, I'm now well over the stage where any poor control usability could be put down to a lack of familiarity - using that touch screen with my left hand, whilst wearing glasses, just isn't something that's either quick or easy. The recent changes that have made the speed display far less visible, as well as effectively hiding all the light icons behind the driver's left hand, haven't improved things, either.
I suspect as this is a safety aspect that you deal with in your job, you have a much more critical eye than the average end user.
Being an average Joe, yes I think the screen looks nice and it makes sense that the UI is like a smart phone. I’ve not personally had any issue with it and can navigate much quicker via the touch screen than I could with the older idrive type systems.
Don’t get me wrong, there’s plenty that can be improved on the car and fsd in my view is being mid-sold, as it can’t be transferred to a car that will actually be capable of fsd
pdk42
Active Member
In truth, the Model 3 would be a much better car from a HMI perspective if it had some physical controls under the central screen, a display in front of the steering wheel and some more controls on the stalks and steering wheel. This is exactly what the VW ID3/ID4 does, but unfortunately they managed to take the edge off success by using haptic feedback touch buttons on the steering wheel and under the central display rather than real buttons. However, it’s still a better design than the Tesla in terms of keeping eyes on the road.
Thats why I don’t use AP. I’m continually tense, knowing that I might have to intervene in a fraction of a second. AP is supposed to make driving more relaxing but I find it does exactly the opposite.
Having seen a few of the FSD videos I think this will never get fixed.
(a) American roads, by and large, are *much* larger than anything we have here, and
(b) there's almost no traffic - I'm watching videos that say 'busy traffic' and it looks like Sunday morning... so AP just can't cope with what we call normal/busy traffic - it's just not coded for it.
I'm not sure I hold out much hope for FSD either if it's being written to cope with those conditions.
(a) American roads, by and large, are *much* larger than anything we have here, and
(b) there's almost no traffic - I'm watching videos that say 'busy traffic' and it looks like Sunday morning... so AP just can't cope with what we call normal/busy traffic - it's just not coded for it.
I'm not sure I hold out much hope for FSD either if it's being written to cope with those conditions.
I have absolutely no confidence that Tesla can achieve FSD on British roads with the current hardware and new beta software. I’d love to be proved wrong but I really don’t think I will be. Achievable on limited roads such as motorways, but it won’t be able to cope with the complexity of city environments.
At least, as far as I know, the Musk hasn’t made any more impossible to achieve predictions recently. He’s been so disastrously wrong in the past I don’t know why people give him any credibility.
Similar threads
