Apple Crash Detection (and how it impacts digital crash investigation)

Apple Crash Detection (and how it impacts digital crash investigations)

On September 7th, 2022, Apple announced the introduction of ‘Crash Detection’ on its forthcoming releases of Apple Watch Series 8 and Apple Watch Ultra, as well as the iPhone 14, iPhone 14 Plus, iPhone 14 Pro, and iPhone 14 Pro Max.

What is Crash Detection?

The feature is reported to detect if you are in a severe car crash where it will automatically connect you with emergency services, provide your location, and notify emergency contacts. It is also reported that Crash Detection runs only when you're driving and processes data only at the time of the crash, entirely on your device.

How Does Crash Detection Work?

Apple Watch Users

In order to tell if you've been in a car crash, Apple said it developed two new motion sensors for Apple Watch Series 8: An improved 3-axis gyroscope and a high g-force accelerometer. These can detect up to 256g force, which Apple said allows it to detect the extreme impact of a crash. These sensors can sample motion up to 4x faster than their predecessors, so Apple Watch Series 8 can sense the precise moment of impact. Apple also built an advanced "sensor fusion" algorithm to enable accurate crash detection. The company studied, for years, vehicle impacts at state-of-the-art crash test labs.

iPhone Users

Similar to the Apple Watch Series 8, the new iPhone 14 models feature new hardware that can detect 256g force. They have a high-dynamic range gyroscope and a new dual-core accelerometer. With these, your phone can automatically notify emergency services and your emergency contacts if you are in a severe car crash. If you have a new Apple Watch and a new iPhone, they seamlessly work together to get you help, Apple said.

What Accidents can Crash Detection Identify?

During its testing and development,?Apple said it focused on four types of severe car crashes: Front-impact, side-impact, rear-end collision, and rollovers.

In each crash test, Apple captured data through its gyroscopes and accelerometers - as well as other Apple Watch Series 8 sensors such as the barometer, microphone, and GPS - and then it used machine learning to incorporate all of these signals into an algorithm that was trained on over 1,000,000 hours of real-world driving and crashes. The result is Crash Detection, which Apple hopes you'll never have to use.?

Which Cars support Crash Detection?

Apple said Crash Detection can work across most popular types of vehicles, like passenger cars, SUVs, and pickup trucks.

How does this affect Crash Investigation?

Crash investigators need to be mindful of this new feature for a variety of reasons as this can affect serious collision investigation in a variety of ways as outlined below:

Additional Crash Data?

It is reported that the feature is utilising data from within the device, whichever that may be iPhone or Apple Watch. It is said that the data is not stored on apple servers but processed on the device itself for the feature to trigger (and is only active when the device is detected as driving).?

It is unclear what the thresholds are for the feature to be activated, i.e. is it once a certain amount of g is reached in a given direction? The High g- force accelerometer is said to be capable of detecting up to 256g’s of impact. We can surmise the trigger is likely to be a certain threshold being reached from the information supplied. This could be from the X or Y plane or because of a rollover. The data is said to be monitored at 3,000 records a second. This is some processing and monitoring from a device as small as a watch when we compare to how often other crash detection systems monitor data and the size of those devices on vehicles.

Testing is reported to have taken place for this feature to be developed and in any type of product development one cannot help but think that this data must have been recorded for the feature to be refined and available for launch which sets out to help save lives. Whether this data resides on the devices or not will be subject to future investigations by digital specialists who I am sure will be itching to get their hands on a device that has triggered. Whether the data is there and accessible is another matter altogether, but we need to be mindful that it may be present and we await the outcomes of those investigations.

If the data is available then a whole host of other factors will come into play over its reliability and as with other crash data it may not be taken on face value but warrant further investigation.

eCall – A disruptive feature?

The Crash Detection feature means that if a crash is detected then the device will call the emergency services and notify emergency contacts. This could be viewed as a disruptive technology. Apple have realised that it was possible to transfer the eCall system to a watch or phone and they appear to have done this. Does this mean that those wishing to have eCall as a safety feature do not need to trade their old car in for the latest model. Effectively anybody with one of these devices, passenger or driver could be upgrading the classic car to have eCall - will new vehicle sales be affected at all?

The fact somebody has an iPhone 14 with them in a vehicle at the time of a serious crash means that other investigative opportunities may exist such as obtaining recordings of what was being said in the vehicle immediately following the crash from phone recordings at the place where the emergency call is being handled. This has been an invaluable tool in the past for me as an investigator of serious collisions, just from drivers reporting collisions to the Police has yielded important nuggets of infromation (suspects are at their most vulnerable immediately following the commission of an offence and often things are said in these conversations that are of significance to an investigation).

The fact that emergency contacts are notified in addition to the emergency services can have significant consequences for Police officers / Family Liaison Offciers. If somebody dies in a collision then the next of kin may already be aware of the crash prior to the police arriving in person to break the news. More switched on criminals may have other people set up as emergency contacts who could also be notified that somebody is incapacitated in the event of a crash which can have implications on serious organised crime.

Artificial Intelligence

Despite Apple saying the data does not go back to their servers are we sure this is not the case. With over 1,000,000 hours of driving to get this feature developed just imagine how more refined the system could be with all users of these devices feeding the AI machine. A million hours of driving is not enough to see how the weird and wonderful ways in which vehicles crash and the more data they have the better the system could become. I will leave this with you for consideration but it would certainly make sense to have this feedback than not for future development of the system. Perhaps those investigating health apps may have some comment here.

Summary

Crash Detection is a significant feature on new technology and it will be interesting to see how investigators will be able to utilise this going forward in their investigations and how the crash community go about carrying out some much needed RnD for this to be fully understood.

Ultimately this will save lives which can only be a good thing.