<![CDATA[Home security is often disrupted by false alarms because conventional systems rely solely on binary logic that does not consider the context of time and the magnitude of window opening. This study designs and implements an Internet of Things (IoT)-based window detection and monitoring system that integrates an MC 38 magnetic sensor and a HY SRF05 ultrasonic sensor, with inference processing using Mamdani Fuzzy Logic on a NodeMCU ESP8266 microcontroller. The system is equipped with a DS3231 RTC module and an NTP synchronization mechanism to maintain timeliness, and provides adaptive responses through LED indicators, buzzer sound patterns, Telegram notifications, and a Flutter-based mobile application. The research objective is to produce contextual alarm decisions (Safe, Alert, Danger) to reduce false alarms without sacrificing response speed. The main contribution is the implementation of a time-aware multi-sensor approach and edge processing so that the system is able to assess the level of urgency based on the physical status of the window, the distance of damage, and the time of the incident. Testing was carried out in tightly closed scenarios, small edits during the day, wide edits at night, and disturbances due to wind or vibration. Test results showed a resolution accuracy of 93.85%, an average ultrasonic measurement error of 0.63% (a difference of <0.5 cm at the test distance), and an average notification latency to the app and Telegram of around 5 seconds. These findings demonstrate that the integration of redundant sensors with fuzzy inference improves intrusion detection evidence in smart home windows]]>
