Identification & Background

Acronym: FERTiOLiVE

Project code | NORTE2030-FEDER-02969400

Project description: The FERTiOLiVE project proposes the development of a disruptive technological solution that is accessible and tailored to the reality of traditional Portuguese olive groves. At the national level, this production system accounts for approximately 75% of the total olive-growing area and is predominantly composed of small- and medium-sized farms, often located in low-density territories with limited access to digital technologies and a lack of effective decision-support tools, particularly in fertilization management. In practice, fertilization is largely based on empirical approaches that are poorly aligned with the actual nutritional requirements of plants, leading to resource waste, productivity losses, and undesirable environmental impacts.

To promote more efficient, informed, and sustainable fertilization management, the project proposes the development and validation of an intelligent and autonomous system integrating a low-cost soil sensor, machine learning models optimized for offline operation on TinyML microcontrollers, and an intuitive mobile application accessible to users with low digital literacy. This solution will enable in-field collection of soil data, inference of plant nutritional status, and the generation of personalized fertilization recommendations in real time and without the need for connectivity, addressing existing technological gaps in the current market.

Approval date |28/09/2025
Start date |01/01/2026
Completion date |31/12/2028

Total eligible investment |1 175 483,25
Financial support | 956 990,12

Promoters | Natureza Prima – Soluções Sustentáveis, Lda
Copromotors | MORE, IPB, APPITAD

  • Development of a plant nutritional status prediction model (soil → leaf), based on real analytical data, reducing reliance on laboratory-based foliar analyses;
  • Creation of an intelligent fertilization recommendation model, trained on legacy data and field sampling, capable of operating locally;
  • Construction of a functional, energy-autonomous sensing prototype, calibrated for relevant parameters (pH, EC, N, P, K, moisture);
  • Integration of the entire system (hardware + software) into a practical tool, validated in real-world conditions, delivering localized and easily interpretable recommendations for growers.
  • A1 – Preliminary Studies and Requirements Gathering;

  • A2 – Development of the Sensing Prototype;

  • A3 – Data Collection, Calibration, and Formulation of Fertilization Rules;

  • A4 – Development of the Intelligent System and AI Models;

  • A5 – Testing and Validation in a Real Environment;
  • A6 – Promotion, Communication, and Dissemination of Results;

  • A7 – Scientific and Technical Coordination and Knowledge Management.

  • Machine Learning model for predicting nutritional status (soil → leaf);
  • ML model for personalized, real-time fertilization recommendation;
  • Low-cost sensing prototype with a TinyML microcontroller, operable in offline mode;
  • Complete and accessible intelligent system (hardware + software).