Fall detection

In Chapter 6, Smart Wearable, we gathered three axis values from the accelerometer. Now, we are going to make use of this data to detect falls.

I would recommend watching the video Accelerometer in Freefall (https://www.youtube.com/watch?v=-om0eTXsgnY), which explains how an accelerometer behaves both when it is stationary and in motion.

Now that we understand the basic concept of fall detection, let's talk about our specific use case.

The biggest challenge in fall detection is to distinguish falling from other activities, such as running and jumping. In this chapter, we are going to keep things simple and work on very basic conditions, where a user at rest or in constant motion suddenly falls down.

To identify whether the user has fallen down, we use the signal magnitude vector or SMV. SMV is the root mean square of the values of the three axes. That is:

If we start plotting the SMV over Time for a user who is standing idle and then falls down, we will end up with a graph, as...