CareConnect: AI-Driven Hospital Environment Monitoring System
CareConnect is an AI-driven system for hospital environment monitoring. Clinical staff can query sensor data using plain natural language, receive real-time visualisations, and trigger automated responses — all through a conversational interface powered by LangChain and large language models.
Motivation
Hospital environments generate continuous streams of sensor data (temperature, humidity, CO₂, occupancy) from wards, operating theatres, and ICUs. Extracting actionable insights from this data typically requires querying databases or reading dashboards. CareConnect replaces these interfaces with a natural-language chat that understands clinical context.
Architecture
- Sensor layer: MQTT broker receives real-time telemetry from environmental IoT devices deployed in hospital rooms.
- LangChain agent: an LLM-powered agent with tool access — it can query the sensor database, request time-series plots, and issue MQTT commands to actuators (e.g., HVAC adjustments, alert notifications).
- Chat interface: staff interact via a simple chat UI; the agent interprets queries, fetches relevant data, and responds with text + charts.
- Automated actions: threshold violations trigger pre-configured actions (notifications, device commands) without human intervention.
Example Interactions
- “What was the average temperature in Ward 3 yesterday?” → agent queries DB and returns summary + plot.
- “Alert me if CO₂ in Operating Theatre 2 exceeds 1000 ppm.” → agent registers a live trigger.
- “Turn on ventilation in Room 14.” → agent publishes an MQTT command.
Technology
Python, LangChain, OpenAI API, MQTT (Paho), Matplotlib, Robot Framework for integration testing.