AI Patient Monitoring System (Vitals + Fall Detection)

Problem statement

In hospitals and home care, continuous monitoring of elderly or critical patients is challenging; falls and abnormal vitals may go unnoticed.

Abstract

This system combines a wearable sensor node measuring heart rate and motion with a bed-side processor running AI models. The node sends vitals and accelerometer data wirelessly. The processor detects falls from IMU patterns and checks for abnormal vital ranges, sending alerts via SMS/app to caregivers.

Components required

• Wearable microcontroller (ESP32 / nRF52)
• Pulse and temperature sensors
• IMU/accelerometer sensor
• Gateway device (Raspberry Pi / PC)
• Wi-Fi/BLE communication
• SMS/app notification service

Block diagram

Working

The wearable node periodically measures vitals and streams IMU readings. The gateway runs scripts that analyze vitals for out-of-range values and classify IMU sequences using a trained model to detect falls. When an event occurs, the system records data and sends alerts to registered mobile numbers or a nursing station dashboard.

Applications

• Elderly care monitoring
• Hospital step-down units
• Rehabilitation centers
• Base for medical IoT products