Why is the sensor showing an incorrect icon?

Sensor is showing an incorrect icon

Of course, this is something we are always working to improve, but from time to time the sensor may have difficulty understanding what is happening in a room. Here are some steps you can take to help us eliminate these problems in the future and some context about why it can sometimes happen.

Reporting incorrect detections

We always appreciate you letting us know if you experience a problem. If you write to us in the support chat and describe the situation in detail (especially who it happened with, when, and what the correct event was), we can get back to you and keep you updated while we investigate the situation.

Possible causes of incorrect detections from the sensor

The sensor sees a room or situation it hasn't seen before

It can be a bit complicated to explain, but for our sensor to understand something in the world, it needs to have seen something similar before. If it experiences a new room or a new situation, it will be more difficult for it to understand the situation unfolding in the room.

For example, if the sensors have only recently been set up at your facility, they might not be fully accustomed to understanding the environment yet.

This can also be the case if a resident has something unusual in their room; for instance, we have previously had to train the sensor not to be confused by pets or large stuffed animals.

The resident is very difficult to identify, for example, if they are almost out of sight

Another situation could arise if a resident would be difficult to see even if you were in the room yourself. This could be if they are in a living room or bathroom and therefore only partially visible, or they are lying in their bed but covered by a large duvet. This is something we are constantly working to improve, and if you tell us when it happens, we can train the system to better understand that situation in the future.

A brief introduction to how our sensor understands the room

Everything the camera sees is automatically understood by the computer and translated into stick figures. These images and texts are what we send on to you.

For the computer to understand what it sees, it must have seen the situation before. You can compare it to a baby needing to see many cats and dogs before it can begin to understand the difference between them. So if we want the sensor to understand what it means when someone kneels on the floor, we need to give it many examples of people kneeling on the floor. The more times the sensor sees the situation, the more confident it will be in understanding what is happening.

That's why we often ask for consent from the residents when we set up at a new nursing home or hospital, as it gives us the opportunity to let the sensor learn about the rooms and ensure that it can understand them correctly.