We are currently experimenting with sound/noise level sensors in our smart home to see if they work effectively and provide a useful capability. We are using a cheap 'off the shelf' sensor, connected to our contextual smart home using an Arduino.
Sound level sensors are fine for detecting generic noise levels but they are not designed to characterise sounds or detect specific sounds. So you can't use these types of sensors to detect dogs barking or glass breaking. This kind of detection requires some frequency filtering and analysis and there are glass break sensors available.
We use common models for all the sensors and all the types of sensors connected to our Home Control System. We have a model for a 'Noise' sensor and it can be used as a trigger for many things. Combined with technology abstraction this means that any noise sensor added inherits all of the capabilities supported by our model and our smart home. Essentially, every sensor connected becomes instantly very smart. They can all generate or voice announcements if required or trigger devices or the alarm.
As currently set up, each noise level sensor is associated with a zone and reports sound levels as a percentage.
We are currently investigating the possible use cases. We have one sensor that is currently being used to detect thunder and also provides an indication of how close it is to our home. We are also looking at how well these cheap sensors can be used to detect activity and be used to trigger things like external lighting and security cameras.
The code we have developed is working really well and we are now testing the sensor at various locations outside of our home.
