Alarm automation strategy for arming

I am in the process of implementing an alarm system using motion sensors, IP cameras, Hue lighting (inside and outside), zigbee switches and HEOS speakers I already possess.

My HA implementation in general is automation-driven rather than dashboard-driven. I don’t want a manual keypad on the wall I have to interact with. Been pondering about a sound way to automate arm(/disarm). The issue is being able to automate arm without risk of false alarms waking up the neighbours or similar.

Current presence detection is a mix of GPS based (app) and Wifi based through DD-WRT integration. DD-WRT seems to be most reliable.

Ideas so far:

  1. Attempt to better failsafe the presence detection through verifying phone status. Example: when presence based on phones is away: verify that phone is online and functioning after time of exiting building. Start timer at trigger to away. Verify that 1. value X (Not sure what value to use. Could use battery % or something similar updated in HA.) change during X minutes, Wifi is on and battery > X%.
  2. Automate arm/disarm, but don’t trigger the alarm using either sound or lights automatically. Only send a loud alert to phone with the option of easily activating alarm. (Video feed and all motion sensors easily available for verification. And it couldn’t be caused by phone hang at home.)
  3. Ditch the automation idea. Use a zigbee switch similar to what’s used for lights to arm/disarm. Not a best security practice, but I suppose using dim up/down or similar would provide some obscurification.

Tried searching without finding any good examples of how to avoid false positives of a fully automated alarm system.

Any ideas or pointers appreciated.

1 Like