Calculate Signal vector magnitude (SVM)

Hello, I'm using the signal vector magnitude formula (SVM) to detect people falling through an accelerometer:

SVM = sqrt (x² + y² + z²);

As I am developing a bracelet, I was unable to disregard the gravity value (9.8m / s²) on the axes. I wonder if this is a problem for the application.

I am making an algorithm that when it detects a high acceleration peak through the SVM equation, consider it as a fall. However, as it is a bracelet, the acceleration of gravity is present in the axes in several ways. Is it a mistake not to remove this gravitational value?

Thank you!

DupTriplicate post.

...and other advice preceded those:

As measured by an accelerometer, free fall is characterized by near zero total acceleration, including the absence of a "gravity" component.

Many modern accelerometer chips have a "free fall" interrupt (to detect a dropped device, etc.).

However, a human being doesn’t really free fall in a slip accident… Imagine what the wrist motions are when you sense that you’re falling to the ground. Think, “flailing around madly”. Also a lot of everyday activities might look like a fall if you don’t know the directions… like closing a car door firmly.