Smart Home Up Up

This article and the developments behind it were inspired by Micky Flanagan's comedy sketch about going 'out out'. It is very funny and helps explain what we are talking about:

The related challenge we have with our smart home is to know if someone is 'properly up' each morning, or as Micky would say 'up up'. It is easy to detect when someone has got up in the night to get a drink from the kitchen or go to the toilet but, we don't want the house to decide that people are now 'up up' and start doing things like running scenes, voice announcements, opening curtains, switching off alarms, unlocking doors, etc. There are many things that we only want to occur in our smart home when it has determined that someone is 'up up'.

To address this we have spent a lot of time refining our smart home's 'up up' algorithm based on detected occupancy and presence, people counting, time of day, event counting and other methods so that we have an accurate view of whether someone in our home has got up or whether someone in our home is 'up up'.

Algorithm

The algorithm used only works because our Home Control System (HCS) has 'whole house' context and monitors occupancy and presence across each zone/room, as well as the whole house.

The algorithm is particularly applicable to telecare and assisted living. The assisted living demonstrator we have built around our smart home software is able to provide a clear view to both local and remote carers as to whether the person being cared for has got up that day and when. If the person being cared for hasn't got up up by a particular time then it will alert carers.

Our algorithm looks for important events that are indicative of the household being awake and assigns them different weightings based on their significance. It then maintains a view of recent activity along with its significance and once this crosses a predetermined 'threshold', it will assume we are up up.

Other Inputs

A key input into this algorithm is provided by bed occupancy sensors connected to our smart home. We have developed our own sensor which provides a very accurate and timely view as to whether a bed is occupied or not. Sensors like these are key to getting quality occupancy and presence data in rooms where people sit or lie still for long periods of time.

We are also monitoring appliance usage in our smart home (e.g. kettle, dishwasher, coffee maker, tumble dryer, etc.). These are also an important input into our algorithm.

Updates

We are currently testing our approach. The algorithm invokes a scenes which also sends a text message informing up that the houshold is up. This allows us to test its accuracy against what is actually happening in opur home. So far it appears to be amazingly accurate, deciding we are up within just a few minutes of first detected activity but, also ignoring the transient events that sometimes occur.

Share ...
We are on ...
Facebook Twitter
YouTube Flickr Follow us on Pinterest