The False Alarm Problem in Home Security
Most security systems suffer overwhelming false alarm rates — 94–99% of all alerts are false alarms (Bureau of Justice Statistics). Primary causes: pets, sunlight reflecting off glass, curtains moving in wind, heat from appliances. A good system must address this first.
PIR Sensor (Passive Infrared)
Principle: detects changes in infrared radiation (heat) as a human body (37°C) moves across the detection zone. Patterns range from curtain-style to wide-angle 90–110°.
Strengths: low cost (50–300 THB), minimal power (<1mA), no RF interference.
Weaknesses: detects transversal movement only (not radial — cannot detect someone walking straight toward the sensor). Bangkok room temperatures of 35°C reduce sensitivity. Pets >15kg can trigger false alarms.
Pet-Immune PIR: the detection pattern is designed to sense only above 50–60cm from the floor. Average Thai housepets below that height (most dogs are <40cm tall) will not trigger the alarm.
Microwave Sensor (Active Doppler Radar)
Principle: transmits 5.8GHz or 10.525GHz microwave pulses and detects Doppler shift in the reflected signal. Detects movement in all directions — both radial and transversal.
Strengths: penetrates glass, thin walls, and doors — can detect someone before they enter a room. Unaffected by room temperature.
Weaknesses: false alarms from wind, fans, air conditioners, and moving plants. 5.8GHz operation may interfere with Wi-Fi. Higher power consumption than PIR.
Dual-Technology Sensor (PIR + Microwave)
Both PIR and microwave must trigger simultaneously (AND logic) to raise an alarm. This reduces false alarms by >90% compared to a single PIR sensor — ideal for homes with pets, fans, or near busy roads.
Recommended for Thailand: Bosch DS150i Dual-Tech (PIR + microwave, pet-immune to 25kg) or Optex VXS-ST (solar-powered for outdoor use).
Temperature Compensation
Bangkok-specific problem: daytime room temperatures of 30–35°C approach body temperature, reducing standard PIR sensitivity. Sensors with temperature compensation (Bosch, Optex) automatically adjust the detection threshold with ambient temperature, maintaining >98% detection rate even in hot rooms.
ML-Based Anomaly Detection in Home Assistant
Home Assistant with HACS anomaly integration learns normal movement patterns in the home (wake times, sleep times, room occupancy) and alerts only on anomalies — such as movement in a room that should be empty at 03:00. This reduces false alarms from routine household movement.
Self-Monitoring vs Central Station Monitoring
Self-monitoring (Home Assistant + LINE OA): free but requires you to receive and respond to alerts yourself — may be missed without mobile signal. Central Station (ISS, ALERT): 300–800 THB/month with a 24/7 operator who contacts police or a security company directly. Best for homes left unoccupied for extended periods.