[sigmabody]

So I recently got prompted for, and updated to, the 03/2023 version of the iDrive software. Since then, I'm getting about 25%+ false positive interventions with emergency braking while backing out of my garage (whereas previously I'd had less than 5 in 1.5 years of ownership, I've had at least five in the last two weeks). It seems to be the same issue which the BMW detection systems have always had (ie: sharp light or color changes cause the system to think something is there, sometimes to very dangerous effect, such as swerving into other lanes at speed), but it seems far more pronounced now.

I also had an random false-positive side impact detection since the update, where the X5 nearly swerved itself into another vehicle.

Has anyone else experienced this since the update? I'm considering taking the vehicle to the dealer to have the system looked at, and possibly asking to revert the software version, since this progressively worse problem seems like it's becoming a safety hazard now. It's very frustrating that this issue has been present since the vehicle was new, and BMW seems to be making it worse with updates.

[shelbymaize]

Since my update I have had no issues at all

[Rob_G77]

Originally Posted by sigmabody

I also had an random false-positive side impact detection since the update, where the X5 nearly swerved itself into another vehicle.

I've been having this issue on my M850i since I bought it a year ago. Lately it's been happening more often and always happens from the right side. Since the dealership hasn't been able to reproduce it themself, they won't/can't do anything. So now I am asking for a case to be opened.

Yesterday on the highway I tried to merge right into another lane and at the same time the collision avoidance tried swerving the car left. I'm very getting tired of this happening... It happens once every day or two. Sometimes twice in a day or twice on the same driving trip.

I think the backup sensing is more sensitive too. Just don't backup so fast and you're ok.

Rob

[sigmabody]

I've now disabled the pedestrian detection entirely. FWIW, backup speed seems to have nothing to do with the issue: I can back up very slowly, and it will still trigger. It seems to be very dependent on sharp lighting differences, though: in the cases of false-positive braking and dangerous false-positive swerving, almost all of them (in my recollection) could be correlated with sharp lighting gradients, either due to reflections of lights at night, or sharp shadows near the detection zone.

If you have a chance to open/escalate a case (ideally with the NTSB, at least in the US), feel free to mention the sharp light/shadows thing as something to look closely at. I feel like I could probably reproduce it artificially using that mechanism, at least based on my experience.