Developing a Low-Cost IoT Prototype for Non-Invasive Elderly Monitoring

An increasing number of older adults spend long periods without direct supervision, increasing the risk of delayed responses to domestic emergencies. While camera-based monitoring systems have emerged as an obvious way to keep them watched, they introduce significant privacy concerns, especially in private spaces. To address this issue, we developed a low-cost IoT prototype called Radar-Cuidador. It provides non-invasive home monitoring by relying on simple sensors and cloud-based alert logic instead of video capture. The hardware design follows a distributed Hub-and-Spoke architecture, using an ESP32 serves as the main hub, integrating an HLK-LD2410C mmWave radar to measure presence, a MAX4466 microphone to capture acoustic anomalies, a BME280 environmental sensor, while for critical areas, an ESP8266 node acts as a spoke, using a Reed Switch to monitor door activity. The system triggers local hardware alerts and transmits data via Wi-Fi to ThingSpeak for automatic processing. Instead of complex image analysis, the alert logic relies on predefined thresholds programmed in MATLAB. During our initial tests, we deployed the hub in a living room and the spoke in a bathroom. This two-zone setup successfully validated the core functionality of detecting presence, logging environmental data, and identifying acoustic events like strong impacts. A complete home deployment would require the implementation of more nodes, however, this prototype proves you can build a working assisted-living system without sacrificing user privacy.