WALADI: Wireless AI Live Alert Device for Infants

Caring for infants, especially newborns, requires continuous supervision to ensure their safety, comfort, and healthy development. However, parents and caregivers often face difficulties maintaining constant monitoring due to demanding schedules, sleep deprivation, and the need to manage several responsibilities at once. Delayed responses to unsafe sleep postures, abnormal breathing patterns, elevated temperatures, or prolonged crying may increase risks to an infant’s health. Traditional baby monitors require caregivers to remain actively attentive. Therefore, there is a need for a smarter monitoring system capable of detecting important changes in the baby’s condition and notifying caregivers in real time.

WALADI is a smart baby monitoring system that combines multi-modal sensing, edge computing, artificial intelligence, and secure IoT communication to support safer infant monitoring. The system is built around a Raspberry Pi 5 connected to several sensing modules assembled in a 3D-printed sensing box that can be mounted on a crib. These sensors consist of a Pi Camera, a thermal camera, an SHT31 temperature and humidity sensor, an mmWave radar, and a microphone. WALADI allows caregivers to monitor the baby’s posture, sleep state, crying events, breathing rate, heart rate, body temperature estimate, and surrounding environmental conditions. A dedicated mobile application was also implemented to allow caregivers to pair their device, view live monitoring values, receive alerts, and access historical data.

Testing was performed at both component and system levels. Individual sensors were first tested separately to verify the quality, consistency, and stability of their readings. Afterward, integrated testing was conducted using realistic baby-monitoring scenarios involving posture detection, sleep state recognition, crying detection, temperature monitoring, and vital-sign estimation. To simulate a real infant safely, the prototype was tested using a baby doll placed under the monitoring device. For thermal monitoring, warm water was circulated through small water pipes positioned near or inside the doll to imitate body heat and allow the thermal camera to detect temperature variations in a more realistic way. These tests helped validate how the system reacts to different monitored conditions without involving an actual infant. The results showed that the implemented prototype can collect real-time data, process it locally, transmit live updates to the mobile application, store historical records, and generate alerts when abnormal or unsafe conditions are detected.

Overall, WALADI provides a functional smart baby monitoring prototype that enhances traditional infant monitoring by combining real-time sensing, edge-based AI analysis, secure communication, and mobile alerting. By helping caregivers respond more quickly to important changes in the baby’s condition, the system aims to reduce caregiver stress, improve infant safety during sleep, and contribute to the broader goals of good health, well-being, and responsible technological innovation.

Project Details

[photo]