The project work plan comprises four phases in consistence with project methodology: 1) project set-up and requirements specification, 2) edge and platform AI development, 3) solution integration and human-machine interaction, and 4) project validation and ethical considerations; aligned with eight Work Packages (WP).

AIPROFICIENT is 36 months long and the objectives will be reached by: one WP for use case characterization and project requirements specification; two WPs on development of edge and platform AI technology, one WP on human-machine interaction and interfaces, one WP on platform integration and interoperability, one WP for use case validation and ethical considerations, one WP for dissemination, exploitation and standardization, and project management WP.

An overview of the project activities is presented as follows:

WP1 – Pilot site characterization, requirements and system architecture

This Work Package delivers the project foundation in the form of ethics by design platform requirements and specification, technical characterization of the pilot sites, operation scenarios and KPIs, an analysis of the operation conditions, legal and ethical issues, and identification of demo scenarios. It will prepare the consortium, and set the conditions for the rest of the WPs.

WP2 – Smart components and local AI at system edge

This Work Package is responsible for the introduction of local AI at the system edge by embedding it at the component level. It will deliver the raw data pre-processing, self-diagnostics algorithms, and component-oriented proactive maintenance, machine vision analysis, but also the control algorithms for closing the control loop from the system edge.

WP3 – Platform AI analytics and decision-making support

This Work Package delivers system-level AI services to the cloud-based platform, and includes development of digital twins, predictive AI analytics, and holistic generative optimisation to meet predefined goals. It will leverage physical modelling, rule-based stream processing, and data-driven AI techniques to develop a holistic multi-objective optimisation approach for improved production planning and execution.

WP4 Human-machine interfaces, explainable AI and shop-floor feedback

This Work Package is aimed to improve the human-machine interaction by developing the interfaces for the connected worker. This will include the development of a plant management dashboard for plant operators and smart mobile app for the workers on the shop floor. WP4 will be also aimed to improve the explainability and transparency of AI decision making.

WP5 – AI-PROFICIENT system integration and deployment

This Work Package deploys the system middleware with gateway transformation services, and integrates the AI services and enabling components (WP2 and WP3) to support the devised methodology. It will perform the integration with the plant automation systems and plant management solutions, as well as the integration with AI enablers of AI4EU platform.

WP6 – Use case evaluation and ethical considerations

This Work Package carries out the demonstration and validation activities to evaluate, through short- and long-term tests, the effectiveness of the delivered solution across all project pilots. It will deal with ethical aspects of the project, perform the instantiation of HLEG guidelines, and provide the recommendations for similar activities, and report on the lessons learnt.

WP7 – Dissemination, exploitation and standardization

This Work Package covers the communication and dissemination of the project results, aiming to increase the visibility of the project, and raise interest in the AI-PROFICIENT solution with the potential stakeholders. It will also develop the preliminary exploitation plan and IPR management strategies, as well as address the standardization related activities.

WP8 – Project management

This Work Package covers project management related activities, including financial, technical, communication, and risk management; as well as quality assurance of the project deliverables (e.g., reports, software, experimental results, guidelines, etc.). It will also set the data and knowledge management plan of the project.