There’s a lot of technology crammed into our Perfect Alert Fall Detection Watch, but perhaps the most important element is the part responsible for deciding that a fall has happened. It’s this central piece of technology that kickstarts all of the other operations, from notifying emergency services and loved ones, to sharing the precise location of the wearer.
To perfect our technology, we have to consider two distinct elements:
- What a fall ‘looks’ like to a computer, and
- How to check that an actual fall has happened (nobody wants us sending alerts over a false signal)
So let’s look at these two parts separately.
How Our Watch Monitors for a Potential Fall
There are three different elements of fall detection technology which all come together inside our slim and lightweight watch design:
- Accelerometers and Gyroscopes: The watch is equipped with these sensors to measure acceleration, orientation, and movement. Accelerometers track changes in speed and direction, while gyroscopes measure rotational motion. Together, they help detect sudden shifts or jolts that might indicate a fall.
- Pattern Recognition Algorithms: The data collected by the sensors is analyzed by algorithms that are specifically trained to recognize patterns consistent with a fall. These algorithms consider factors like the intensity of the impact, the direction of movement, and the sudden deceleration that typically occurs when someone falls onto a solid floor.
- Angle Detection: If the watch detects a sudden change in position (like a rapid sideways swing of an arm if wearing a watch), it cross-references this data with the person’s usual movements. Staying in a horizontal position for an extended period can trigger the alert if it follows a suspected fall.
When a fall occurs, the accelerometer picks up a sharp shift in motion, while the gyroscope assesses the angle at which the person’s body moves. These sensors send data to the watch’s algorithm, which processes the information to determine whether the movement is consistent with a fall. Unlike simpler systems that only rely on impact detection, our algorithm helps to differentiate between common daily motions, like sitting down quickly or dropping an object, and actual falls.
You can see how, taking all of this data together, the on-board computer can decide whether it’s likely that the wearer has fallen.
But, before firing the alert, we have to put a layer of buffering in place to make sure we don’t call the cavalry for what’s known as a ‘false positive’, i.e. when the software thinks a fall has happened when it hasn’t.
Ensuring a Real Fall Has Happened
There’s a real balance to strike with monitoring for a false positive. If there has been a fall, we need to call for help as soon as possible but, if there hasn’t, we don’t want to alarm anyone. Essentially, any false positive checks the watch makes have to be carried out quickly and be effective.
Our Perfect Alert Falls Watch takes the following additional measures to counter false positives:
- Ask the wearer to confirm if they’re okay before sending alerts
- Sound an alarm that if not canceled within 25 seconds causes the alerts to be sent
By using sensor data in real-time, the system ensures that users receive fast and reliable alerts, notifying loved ones or emergency contacts almost instantly when a fall is detected.
Sending Out the Emergency Alerts
If a fall is detected and not dismissed by the wearer within a short time, the watch sends an alert to pre-configured contacts with the person’s location information. This feature is especially useful for seniors or individuals with conditions like dementia.
One unique feature of Perfect Alert is that it’s the only watch that can call up to 3 emergency contacts together (Flawless Connect Plus) connecting to the first to answer to save vital minutes.
Learn more about what happens once a fall is detected (link opens a new tab).
The Accuracy of Fall Detection Watches
The accuracy of fall detection in smartwatches has been a significant area of research, especially as these devices become more widely used by seniors and those with mobility challenges. On average, studies indicate that fall detection watches can achieve accuracy rates between 80% and 95%, depending on the model and the specific circumstances surrounding the fall. This number is affected by the age, mobility and behaviour of the user.
Factors That Affect Accuracy
One of the key elements influencing fall detection accuracy is the user’s physical condition. For instance, older adults with slower movements may experience a different pattern of falls compared to younger users or those who move more quickly. As a result, the watch’s sensors and algorithms might interpret certain slow, controlled falls differently than more abrupt movements. User behavior, such as wearing the watch loosely or not wearing it consistently, can also affect its ability to accurately detect falls. Our focus on specifically helping the elderly and their relatives is the reason our Perfect Alert Fall Detection Watch is one of the most accurate available in the UK today.
Differences Between Brands and Models
Not all fall detection watches perform equally. Some watches rely purely on basic sensor data, while others incorporate more advanced algorithms to interpret motion data. The use of algorithms in fall detection, like those found in the watch we offer, enhances accuracy by filtering out non-critical movements. As a result, certain brands may be more prone to false positives, while others might miss detecting certain types of falls, particularly slower or partial falls.
Our watch, for example, uses finely-tuned algorithms that have been optimized for a wide range of movements, helping to reduce the chance of missing a fall. The integration of multiple sensors also improves the overall reliability, making it more likely that a true fall will trigger an alert.
Common False Positives and Negatives
Like any technology, fall detection watches are not perfect. One of the most common issues users face is false positives — instances where the watch incorrectly detects a fall. These can happen during activities such as sitting down abruptly, dropping the watch, or making a rapid movement that mimics the dynamics of a fall. While false positives can be inconvenient, they are generally harmless, as our watch allows users to cancel an alert if no real fall has occurred.
On the other hand, false negatives, where a real fall is missed, can be more concerning. These tend to happen during slow falls or situations where the user may slide or collapse gradually because the triggering sharp movements being monitored for don’t happen. Although fall detection accuracy is improving with each generation of devices, this remains an area where improvements are continually being made. For example, our watch has an alarm button on its face so the wearer can always trigger an alarm manually if they take a fall that the watch doesn’t detect. Happily, at Perfect Alert, we’re proud to have one of the lowest incidences of false positives in the UK.
If you’d like our technology to bring you or your elderly relative peace of mind to know that they’re being looked after 24/7, then click this link to find out more about our Perfect Alert Fall Detection Watch.