I’m working on some home automation SW for a Security usecase I imagine to have ultimately an ontology with different sources of information:

• camera(s)
• microphones
• burglary
• List item
• alarm
• other sensing equipment’s

The events I would like to detect, using a mix of image / sound recognition etc. are:
burglary

• my son coming home with guests (!)
• window break
• fire
• dog barking
• etc.

And ultimately I have a list of actions:

• warn or call police
• call my neighbour
• call my brother that lives in the same street
• warn or call the fire brigade
• call RSPC (animal protection agency)
• etc.

Ultimately I would like to build a mini-reasoning structure able to suggest different actions in relationship to different input, taking into account that false positive exist and need to be avoided (i.e. offering some level of cross-checking).

Potentially this would take place on an edge device of some type.

I would be interested to study similar projects and to get some ideas.

I’m looking at RDFox but I’m concerned it may be too big for my usecase.

Suggestions / ideas welcome!