iPhones and coasters

Interesting:

https://www.cleveland19.com/2022/10/10/iphone-14-reportedly-dials-9...ter-rides/

I work at a dispatch center and can confirm. (eyeroll)

Simple enough fix - exempt amusement parks from crash detection using geolocation. They'll probably have this remedied in no time.

The last anecdote in that article is interesting, where they say the fix "is simple," and to put the phone in airplane mode before going on roller coasters. Is that the guidance from Apple? Or just the writer? And do they really expect every iPhone 14 user to do that?

Honestly that seems like a pretty significant software bug, and I'm surprised they're even allowed to have a feature that automatically calls 9-1-1 on a phone.

It's actually been rather useful by several accounts, so far. I've read stories of a few people who were in crashes and the feature activated - one guy was riding his bicycle and his watch was able to call 911 for him when he crashed pretty hard - and a couple of others who got in bad enough wrecks that their phones flew out of reach in the car and they couldn't get to it.

Android has had this feature for three years. I don't recall it having this problem.

Coasters can be a little jarring and violent. I can’t help but wonder? Was it a woodie? (Beast, Racer, Timbers)

I think you are supposed to wait 4 hours to call a doctor for that.

If Son of Beast were still around, the "erroneous" calls would be warranted.

I remember getting a 911 call from someone on the Millennium because they were stacked too long on the brake run. Nothing medical. Just didn't want to wait...

I will say that the "crash detection" is not the issue. It's hitting the power button 5 times in a row that will cause iPhones to call 911. There are easily at least 100 butt-dials a day from Cedar Point. It does make for a fun game of "Guess the rollercoaster based on their scream!"

Valid point! I once had my phone on my lap at a concert, looked down, and great already connected with 911 “sorry, I clearly don’t know how my phone works all’s well”

Jeff:

Android has had this feature for three years. I don't recall it having this problem.

They've also had 911 since inception yet, once again, the Pixel 6 & 6a can't call emergency services all the time when needed. ;)

Yes, I've read those anecdotes. You can find the same thing about iPhones even before Pixel was a thing. It's easy to blame the phones when we don't know anything about the network that it's connected to.

Another reason to avoid buying Apple products

There's no denying Apple makes fantastic hardware. To me the only reason to avoid Apple is if you don't want to live exclusively in their ecosystem.

Isn't it based a g forces? A simple google search shows the highest g forces for a current roller coaster is 5.2. A 30MPH car crash against a fixed object is in the neighborhood of 30g. Seems a simple tweak to the range of force before it goes off would suffice. What do they have it set at now? 1.5g?

There are a few complicating factors with using G forces. While you're going to see relatively low sustained forces on roller coasters, the impact loads (under 250 milliseconds per the standard) can get pretty high. Add to that, the phone is an unsecured object rattling around loosely coupled to another unsecured object loosely coupled to the ride vehicle (except on Steel Vengeance). As the ride applies its loads, instead of transferring those loads directly to the rider, the ride causes the rider's position to displace until it meets something solid, at which point that body will not just be stopped but will be shoved in the opposite direction based on the relative motions of rider and seat. This is the point where Magnum riders go "Ouch!". The acceleration is an impulse, only a few milliseconds, but it can be a very high magnitude, which is what sets the phone off.

Watches have an even tougher situation in that your wrist can generate pretty extreme forces. I know most of us don't do this on a regular basis, but consider a baseball pitcher who can throw a 100 MPH fastball. To do that, he has to put a force on the ball high enough for it to go 100 MPH. That's...let's call it 146 feet per second. That means the hand has to be moving that fast at release. The arm is about 2' long, so the wind-up can only be about 6'. That means the average acceleration of that fastball during the windup, assuming I am anywhere near getting my math right, is about 55G! It's only for 0.82 seconds, but to go from 0-146 feet per second in a span of 6' requires an average acceleration of 1,776 ft/sec/sec. If that fastballer is wearing a smart watch on his throwing hand, he's likely to set off the crash detector if only the G force is measured.

So it would make sense to use other means of detecting whether this is likely to be a crash scenario. For instance, is the device moving at a speed that suggests a vehicle is involved (on a roller coaster: yes). Is the acceleration combined with a sudden change in altitude and/or orientation? (on a roller coaster: probably). Is the crash event isolated (on a roller coaster: maybe; maybe not)? Maybe location needs to be made more important: is the location consistent with a road, highway, bike path or foot path? Is the crash severity consistent with an expected mode of transportation? With a dropped phone? With an ambulatory fall?

So, yeah, crash forces are not just for crashes. It's not an easy problem to fix, but by combining enough data they can probably fix it.

--Dave Althoff, Jr.

Good points Dave! Also, a rollercoaster doesn't go from very high g forces to 0 in a fraction of a second, like a car crash. Well, I guess the end of Wildcat would be an exception to that. :)