PesTwin: a biology-informed Digital Twin for enabling precision farming
A new AI framework simulates insect invasions by fusing lab biodata, weather, and GIS maps for precision farming.
A multi-institution research team has published a paper on PesTwin, an innovative digital twin framework designed to simulate and forecast pest invasions in agricultural settings. The system is built on a flexible, rule-based Agent-Based Modeling (ABM) paradigm, which allows it to finely tune the complex ecological interactions between a pest, its host crop, and the surrounding environment. By integrating heterogeneous data sources—including precise pest biodata from laboratory studies, real-time environmental data from weather stations, and spatial GIS data of actual crop fields—PesTwin creates a dynamic, predictive model of infestation spread across both space and time.
The framework was specifically applied to model the Spotted Wing Drosophila (Drosophila suzukii), a globally invasive fruit fly that causes significant damage to soft fruit crops. This application demonstrates how PesTwin operates within the principles of precision agriculture and Integrated Pest Management (IPM), moving beyond generic models to provide scenario-specific forecasts. The 6-page study, submitted to arXiv, presents the architecture and includes 5 figures detailing the model's components and outputs, showcasing its potential to become a critical decision-support tool for farmers and agronomists.
- Uses Agent-Based Modeling (ABM) to simulate complex pest-host-environment interactions for accurate forecasting.
- Integrates three key data types: laboratory pest biodata, real-time weather station data, and spatial GIS crop field maps.
- Successfully applied to model the invasive Spotted Wing Drosophila, a major global pest for berries and soft fruits.
Why It Matters
Enables data-driven, targeted pest control, reducing pesticide use and crop loss to boost food security and farm productivity.