Detect Interaction with a Hinged Object

Detect the opening of a door or lid, or manipulation of a handle or tab.

reelyActive Enabling Computers to Understand Who/What is Where/How

The TL;DR (Too Long; Didn't Read)

Learn a simple yet effective means to observe and analyse interaction with any hinged object.


Hinged object?
Doors, lids, handles, tabs — anything that pivots on a hinge with significant motion.
Why detect this?
Occupancy analytics in the case of a door. Operational efficiencies in the case of a trash lid. Compliance in the case of a hand wash station. The list goes on...
Does this really work?
We've certainly been using it successfully at reelyActive Parc!

How this works

The beacon transmits accelerometer data on motion, which is interpreted as an interaction

Configure the sensor beacon   Step 1 of 4

Configure a beacon to advertise when motion is detected by its accelerometer.


Can any beacon be used?
No. Only beacons with an accelerometer will be able to detect when they are in motion.
What beacons have accelerometers?
Such beacons include the Minew E8 and the second-generation Puck.js among others.

Install the sensor beacon   Step 2 of 4

Affix the beacon to the hinged object so as to maximise its motion during an interaction.


Is placement important?
Yes. If there's insufficient motion to trigger the beacon, the interaction will not be detected.
What placement is optimal?
In general, placement farthest from the hinge will produce the most motion.

Calibrate the interaction detection   Step 3 of 4

Determine the threshold of unique sensor readings that correspond with an interaction.


Why unique sensor readings?
When motion is detected, the beacon will periodically transmit real-time accelerometer data, with each data reading likely to be unique during an interaction.
Why not a standard threshold?
A default threshold may indeed produce reliable results in some scenarios, but this is best validated and adjusted on a case-by-case basis.

Create the visualisation in Kibana   Step 4 of 4

Create a time series visualisation annotated with each detected interaction.


Why a time series?
A time series visualisation highlights interaction patterns over the course of a day, week, month or any pertinent timeframe.
Are there alternatives?
Of course! Kibana provides plenty of visualisation options, and Elasticsearch affords a rich set of queries. What we present here is a starting point.

Where to next?

Continue exploring our open architecture and all its applications.