IVACE - FEDER COGNOSFOOD

Objective

The objective of the project is to develop a prototype system based on artificial intelligence/Machine Learning and advanced interfaces (augmented reality, virtual assistants and wearables) around a common cognitive industry architecture that transforms the way food is manufactured so that key industrial processes have their own artificial intelligence and are executed in a coordinated and orchestrated manner through intelligent data analysis.

Through the development of the FOOD COGNITIVE INDUSTRY paradigm, AINIA aims to contribute directly to all lines of action, so that the future transfer of the developments of the COGNOSFOOD project is maximised, in particular so that companies in the Comunitat Valenciana improve, individually and as a whole, their position compared to the rest of the national territory and against competing companies in international markets.

The specific objectives are:

Develop a proposal for a reference architecture for the food cognitive industry in which to locate the different cognitive components based on artificial intelligence, addressing current issues in production, food safety, value chain and innovation.
Develop prototypes of decision-support applications that allow simulations to be carried out to optimise operations in a coordinated way within the organisation, in dimensions such as cost, sustainability and process efficiency from a holistic point of view.
Design a technological structure for massive capture, processing and governance of process data in the plant, characteristics of the food matrix and physicochemical and microbiological laboratory parameters, enabling the acquisition of data with guarantees to be used in the modelling of predictive systems aimed at:
- Predicting the descriptive parameters of the food product (quality parameters) and establishing risk levels for the appearance of microorganisms in the production plant
- Improvement in production processes through the combination of Co-bots and advanced food inspection systems.
Implement a prototype of intelligent technology for controlling the information flow of distributed processes capable of effectively synchronising different key processes (inter-organisations) of the value chain aimed at quality and food safety, and effectively supporting companies and suppliers in the process
Develop an integrated prototype of food cognitive industry of all these subsystems to demonstrate its execution capacity and integration with existing systems in pilot lines
Piloting of the developed prototypes using datasets (data sets taken from production, chain and quality control processes) from benchmark companies in the agri-food sector of the Comunitat Valenciana.

Activities

Design of the systems architecture for cognitive food manufacturing
Research and digital characterisation of cognitive food manufacturing processes
Development and implementation of the technologies and CORE components of the cognitive food manufacturing system
Design and development of user applications and integration of the cognitive food manufacturing system prototype.
Unit testing and piloting with users
Dissemination and communication
Transfer and promotion of results
Management-coordination

Results

New production model based on the concept of Food Cognitive Industry founded on the ubiquitous presence of artificial intelligence, taking into account opportunities for improvement in a company’s key production processes.
Methodologies for generating robust corpora of process parameters as input variables (decision variables) and quality parameters as output variables (dependent variables).
Prototype of a Big-Data core that will manage and contextualise a set of heterogeneous data under conditions of high arrival speed, high variety and variability in a related way, enabling decision-making horizontally across all operations and specifically in each one of them.
Integration of cobots and advanced inspection systems.
Predictive process models created and trained to identify the influence of process parameters on the descriptive parameters of product quality, enabling “What If” simulations to be carried out.