PilotAssist critical flaw

[EssexExile]

Originally Posted by nu11eaf

I had a learning curve using PA, the first lesson learnt was, pulling off a 50mph dual carriageway onto a very long off ramp heading towards a roundabout at the end, and finding the car not slowing as it heads towards the stationary cars waiting at the roundabout and applying last minute heavy braking, car stopped on its own but it was scarey "lesson learnt" stationary traffic requires driver response. If the radar is not locked on to a moving car in front it will not react, in my case it was collision avoidance that stopped me not PA.

I find this very worrying. Were you just going to let it drive into the cars in front if it didn't brake itself? I'd have been on the brake way before "heavy braking" was needed.

[nu11eaf]

Originally Posted by EssexExile

I find this very worrying. Were you just going to let it drive into the cars in front if it didn't brake itself? I'd have been on the brake way before "heavy braking" was needed.

No I had reduced my set speed via the steering wheel button when approaching the roundabout so was not going very fast and was surprised that the car had not started braking and ready to brake when the car pulled up. Had only used PA a few times.

[scw]

In my experience, there's a definite cut-off point in relation to traffic density where monitoring the behaviour of PilotAssist starts to use more cognitive bandwidth than watching the road ahead. At that point, it's time to fall back on adaptive cruise. Even the latter can be a disadvantage in certain traffic scenarios.

On quiet, early morning (5-6am) commutes on the A1/A14, such as I had a couple of years ago, PilotAssist was great. Less so with the usual congestion during normal hours.

I'm not expecting AI to completely take over the driver's responsibility for quite some time; the environment is just too chaotic.

Steve