Skip to main content

EcoTwin Agent

AI-powered environmental health advisor providing personalized recommendations based on real-time air quality, weather forecasts, and predictive dispersion modeling.

Overview​

The EcoTwin Agent acts as a digital twin for urban environmental health monitoring, combining:

  • Gaussian plume dispersion modeling for pollutant spread prediction
  • Google Gemini AI for personalized health advice
  • NGSI-LD entity generation for real-time visualization
  • Citizen report integration for crowdsourced validation
graph TD
    A[EcoTwin Agent] --> B[Air Quality Data]
    A --> C[Weather Data]
    A --> D[User Profile]
    A --> E[Gemini AI]
    
    B --> F[Dispersion Model]
    C --> F
    
    F --> G[AQI Forecast]
    
    D --> E
    G --> E
    
    E --> H[Health Recommendations]
    E --> I[Activity Suggestions]
    E --> J[Protective Measures]
    
    H --> K[NGSI-LD Entity]
    I --> K
    J --> K
    
    K --> L[Stellio Context Broker]

Core Capabilities​

1. Air Quality Dispersion Simulation​

  • Gaussian plume model for pollutant dispersion
  • Wind direction and speed integration
  • Temperature inversion layer detection
  • Multi-pollutant tracking (PM2.5, PM10, NO2, O3, CO, SO2)

2. Personalized Health Advice (Gemini Pro AI)​

  • Vulnerable population profiling (elderly, children, asthmatics)
  • Activity recommendations (outdoor exercise, window opening)
  • Protective measure suggestions (masks, air purifiers)
  • Natural language explanations in Vietnamese and English

3. Environmental Forecast Publishing​

  • Future AQI predictions (1-6 hours ahead)
  • NGSI-LD entity generation for visualization
  • Historical trend analysis
  • Confidence scoring

4. Citizen Report Integration​

  • User-submitted health symptoms correlation
  • Crowdsourced air quality validation
  • Community health alerts

API​

generateHealthAdvice​

Generate personalized health recommendations based on location and user profile.

interface HealthAdviceRequest {
    location: { lat: number; lon: number };
    userProfile: {
        age: number;
        conditions: string[];  // e.g., ['asthma', 'heart_disease']
        activities: string[];  // e.g., ['outdoor_exercise', 'commute']
    };
    timeHorizon: number;  // hours (1-6)
}

interface HealthAdviceResponse {
    recommendations: string[];
    aqiForecast: {
        current: number;
        predicted: number[];
        confidence: number;
    };
    riskLevel: 'low' | 'moderate' | 'high' | 'very_high';
    protectiveMeasures: string[];
    aiExplanation: string;
}

const advice = await ecoTwin.generateHealthAdvice(request);

getDispersionForecast​

Calculate pollutant dispersion based on current conditions.

interface DispersionRequest {
    sourceLocation: { lat: number; lon: number };
    pollutant: 'pm25' | 'pm10' | 'no2' | 'o3' | 'co' | 'so2';
    windSpeed: number;
    windDirection: number;  // degrees
    atmosphericStability: 'A' | 'B' | 'C' | 'D' | 'E' | 'F';
}

const forecast = await ecoTwin.getDispersionForecast(request);

AQI Categories​

AQI RangeCategoryHealth ImplicationsColor
0-50GoodSatisfactory air qualityGreen
51-100ModerateAcceptable for mostYellow
101-150Unhealthy for SensitiveRisk for sensitive groupsOrange
151-200UnhealthyHealth effects for everyoneRed
201-300Very UnhealthyHealth warningsPurple
301-500HazardousEmergency conditionsMaroon

Configuration​

Located at config/agents/eco-twin.yaml:

agent:
  name: EcoTwin
  version: 3.0.0
  description: Environmental Health Digital Twin

dispersion:
  model: gaussian_plume
  gridSize: 100  # meters
  predictionHorizon: 6  # hours

ai:
  model: gemini-1.5-flash
  temperature: 0.7
  maxTokens: 2000
  languages: [vi, en]

caching:
  enabled: true
  ttl: 300  # 5 minutes

Dependencies​

  • @google/generative-ai@^0.21.0: Gemini Pro AI SDK
  • axios@^1.6: HTTP client for Stellio API
  • js-yaml@^4.1: Configuration file parsing

Usage Example​

import { EcoTwinAgent } from './agents/EcoTwinAgent';

const ecoTwin = new EcoTwinAgent();

// Generate health advice for elderly person with asthma
const advice = await ecoTwin.generateHealthAdvice({
    location: { lat: 10.8231, lon: 106.6297 },
    userProfile: {
        age: 65,
        conditions: ['asthma'],
        activities: ['morning_walk']
    },
    timeHorizon: 6
});

console.log(advice.recommendations);
// ["Avoid outdoor exercise before 10 AM due to elevated PM2.5",
//  "Wear N95 mask if going outside",
//  "Run air purifier on high setting"]

console.log(advice.aqiForecast.predicted);
// [125, 118, 105, 95, 88, 82]  // Next 6 hours

See Also​