LLM-Based Intent Orchestration for Medical IoT Environments

Overview

Modern medical research laboratories increasingly integrate smart workspace environments with diverse IoT devices and services. However, clinicians, nurses, and researchers—typically non-IT specialists—require intuitive mechanisms to express their operational intents without manual device configuration. LLMs offer promising capabilities in reasoning, planning, and task orchestration, enabling seamless automation of data retrieval, analysis, and workflow execution. More details in CICT Hackathon Presentation

Application Scenarios

Scenario 1: Real-Time Physiological Monitoring

Wearable and implantable sensors continuously monitor vital signs (blood pressure, heart rate) with anomaly detection and immediate caregiver alerts.

Research Direction: Edge-based lightweight patient prediction models (distillation) running directly on sensor devices.

Scenario 2: Smart Hospital BioIOT Management

Camera-based monitoring systems detect resident falls, autonomous medical IOT sensors/actuators management, enhancing safety, performance, energy-awareness while reducing staff workload.

Research Direction: Generative AI multi-agent for task orchestration, deployment, auto-configuration, plan validation and execution.

Quick Setup

Prerequisites

• Install pipx: github.com/pypa/pipx
• Install Poetry: python-poetry.org/docs

Installation

# Verify Poetry installation
poetry --version

# Install dependencies
python -m poetry install (--verbose)

# Check virtual environment
poetry env list

# Activate environment (Poetry 2.x)
source $(poetry env info --path)/bin/activate

# Or for Poetry 1.x
poetry shell

Project Structure

llm-intent-orchestration/
├── src/
│   ├── main.py              # CLI entry point in FastMCP server with menu interface
│   ├── crew.py              # Multi-agent orchestration
│   ├── agents/
│   │   └── agents.py        # LLM Agent logic (Gemini LLM)
│   ├── tasks/               # Task router 
│   ├── mcp/                 # MCP Server tools
│   ├── prompts/             # System prompts
│   ├── utils/               # External task tools
│   └── db/                  # Database
├── configs/                 # Configuration files
├── tests/                   # Unit tests
├── docs/                    # Documentation
├── tools/                   # Tools
├── data/                    # Data
└── .env                     # Environment variables

Running the Application

# Seed database mitigation (Future replacement with real-time iot device data)
python -m poetry run python src.db.main 

# Uvicorn App
python -m poetry run uvicorn src.main:app --host 0.0.0.0 --port 8001

References

1. MCP SDK Integration: modelcontextprotocol.io/docs/sdk
2. CrewAI Task Automation: docs.crewai.com/en/mcp/overview
3. CrewAI Tutorial: youtu.be/sPzc6hMg7So
4. MCP Learning Resources: youtu.be/QIOk4XZ5XNU
5. CrewAI + FastMCP: github.com/ashishpatel26/Crewai-MCP-Course
6. Integration with FastMCP via langchain-mcp-adapters
7. ONOS MCP Server (Code inspiration):onos-mcp-server