A Lightweight (Recurrent Neural Network) RNN Architecture for IoT Based Baby Monitoring System to Provide Real Time Alerts and Notification to Parents or Care Takers

Our modern way of life is characterized by its busy, hectic nature. People are constantly on the move, juggling various responsibilities, which can make it difficult to dedicate uninterrupted care to infants. Working parents are the one facing a major parenting challenge. The paper proposes an IoT device build using light weight Recurrent Neural Network (RNN) architecture that monitors an infant's vital signs, safety, and environmental conditions in real-time, providing parents with real-time insight into the well-being of their child. By using smartphones and other connected devices, parents can access vital information about their babies remotely. Parental confidence is further enhanced by the incorporation of safety features like motion sensors and Accelerometer sensor. If certain conditions are met, the moisture level of the diaper is tracked by the humidity sensor, and the parents are notified by text or phone call. The RNN architecture implemented in our proposed device is capable to recognize the gesture and motion of the baby in the cradle. The paper outlines Lightweight Recurrent Neural Network (RNN) architecture for computationally intensive and memory Constrained IoT Based Bay monitoring system. Results from the prototype were compared with certified market alternatives proving the proposed baby monitoring system's cost-effectiveness and energy efficiency.