“PRAGMATIC” Project

The project “PRAGMATIC (Prediction of Agriculture Manufacturing Costs). R&D work on the prediction of yielding and raw material costs for agricultural products sourced in the supply chain from crop to production line” Funding agreement no. POIR.01.01.01-00-2298/20-00 of November 30, 2021.

Project objective and intended end result:

The goal of the project is to develop a prototype of an innovative IT platform with algorithms for predicting the production costs of agricultural raw materials for three reference crops, i.e. blueberries, apples and potatoes, in the supply chain from field to production line. The system will employ predictive models based on machine learning and artificial intelligence methods, making use of ground data (soil, plant, climate, cultivation technology) and satellite imaging data (crop and soil condition). It will comprise the following modules:

– the Knowledge Base (KB) Module in the form of a data repository, continuously synchronised with available datasets,

– the Violation and Anomaly Detection (VAD) Module, which will make it possible to detect data records that do not match the datasets under review, – the Yield Prediction (YPR) Module, which will enable the prediction of yields for a selected crop type, considering a variety of factors (attributes),

– the Cost Prediction Module (CPR), which will support users in production optimisation and budget planning,

– Performance and Configuration Visualisation Library (PCV). The library will comprise various components used by the analytical modules to visualise performance and display it as diagrams and charts with appropriate labelling of prediction confidence levels (YPR and CPR modules) and signalling of possible problems when anomalies are detected in the data (VAD module).

The project’s main beneficiaries will be medium-sized and large farms growing selected apple, potato and blueberry varieties, which are looking for smart, precise, algorithm-based tools to enable them to produce yields appropriate to market demand and conduct agro-technical operations in a way that is friendly to the environment, the climate and people.

The goal of the project will be achieved through the following actions:

  • developing and producing laboratory hybrid predictive models based on the data most available to the end user, without defining varietal diversity within species;
  • producing tools to conduct the optimisation and testing process in a near-real environment;
  • preparing a prototype for commercialisation by involving end users in the evaluation and validation of the developed system.

Project Value

The total eligible expenditure is PLN 3,753,823.14.

Contribution from the European Funds

The maximum grant amount – contribution from the European Funds – is PLN 2,851,927.94, i.e. 75.97% of the total eligible expenditure.

image