Periodic Reporting for period 1 - FAIRWork (Flexibilization of complex Ecosystems using Democratic AI based Decision Support and Recommendation Systems at Work)
Période du rapport: 2022-09-01 au 2024-02-29
Current automated and hierarchically structured production processes can only insufficiently deal with the upcoming flexibilization. We foster the “democratization” of decision-making in production processes, hence the participation of all involved stakeholders, by introducing a decentralized AI system. Our Democratic AI-based Decision Support System (DAI-DSS) democratically finds the appropriate decision for a concrete situation during production. Each human or technical actor is represented by an agent who negotiates based on the current status provided by the digital shadows and twins. The future situation is predicted by AI algorithms for each individual actor considering the modelled knowledge base that defines each negotiation strategy. A multiple optimisation algorithm finds the most appropriate solution considering the needs of all involved human and technical stakeholders.
FAIRWork brings “Human, AI, Data and Robots” together, (a) by introducing Design Decision Models and assessing if those models are “appropriate / fair”, (b) decentralizes the decision-making, by representing each involved – human or technical – actor within a Multi Agent System and configuring each agent with previously approved decision models, and (c) broadens the view to optimize the production process, by also introducing social and operational-related parameters, like energy consumption, in addition to the existing technical and business-related aspects. These will be identified and evaluated in user-centred decision making.
Our aim is to:
Improve optimization of the production processes. Raise efficiency through faster and more precise decision-making, integrating digital twins from all involved process stakeholders. Raise energy efficiency by configuring process management techniques, so that digital twins concerning energy-efficiency of business processes can be managed. Improve co-creation capabilities and strengthen compliance by using a transparent model-based approach that enables the collaboration with relevant stakeholders during decision-making.
Strengthening the perception of trust and fairness. Improve workforce wellbeing and raise competence via transparent and co-creative configuration of decision-making, encouraging the consideration of the heterogeneous needs of relevant stakeholders. Raising attractiveness for skilled people by introducing novel technologies like AI, digital twins or smart devices.
Improve flexibility of European Industry. Cooperative decision-making has the potential to be extended with the vision that conveyor belt-oriented production moves towards co¬operating robots. Collaborative decision-making can be applied between different actors within a production process including robots and operators. Support companies in offering “low-lot size” production capabilities that are especially relevant for start-ups. Intended time-reduction in prototyping raises the overall flexibility and reduces barriers by improving “low lot size” production.
FAIRWork brings “Human, AI, Data and Robots” together, (a) by introducing Design Decision Models and assessing if those models are “appropriate / fair”, (b) decentralizes the decision-making, by representing each involved – human or technical – actor within a Multi Agent System and configuring each agent with previously approved decision models, and (c) broadens the view to optimize the production process, by also introducing social and operational-related parameters, like energy consumption, in addition to the existing technical and business-related aspects. These will be identified and evaluated in user-centred decision making.
Our aim is to:
Improve optimization of the production processes. Raise efficiency through faster and more precise decision-making, integrating digital twins from all involved process stakeholders. Raise energy efficiency by configuring process management techniques, so that digital twins concerning energy-efficiency of business processes can be managed. Improve co-creation capabilities and strengthen compliance by using a transparent model-based approach that enables the collaboration with relevant stakeholders during decision-making.
Strengthening the perception of trust and fairness. Improve workforce wellbeing and raise competence via transparent and co-creative configuration of decision-making, encouraging the consideration of the heterogeneous needs of relevant stakeholders. Raising attractiveness for skilled people by introducing novel technologies like AI, digital twins or smart devices.
Improve flexibility of European Industry. Cooperative decision-making has the potential to be extended with the vision that conveyor belt-oriented production moves towards co¬operating robots. Collaborative decision-making can be applied between different actors within a production process including robots and operators. Support companies in offering “low-lot size” production capabilities that are especially relevant for start-ups. Intended time-reduction in prototyping raises the overall flexibility and reduces barriers by improving “low lot size” production.
Specification have been constituted from the use cases partners - FLEX “Automated Test Building”, “Worker Allocation”, “Machine Maintenance After Breakdown” and CRF “Workload Balance”, “Delay of Material” and “Quality Issues”. Decision Processes have been modelled requiring support for: “finding similar projects/experts”, “simulate production process”, “allocate / map worker with profiles”, “find similar problems”, “allocate order to production line”, “assess the impact”, which are realized using different AI services. The Knowledge Base is built on open ISO 10303 (STEP) standard for data exchange. The results are published in D2.1 and D5.1.
Relevant literature and developments related to (a) democratization of decision-making and digital twins for human experts, (b) technical approaches like Artificial Intelligence (AI) and Multi-Agent System (MAS), (c) aspects of reliability and trustworthiness in AI is provided. Research has been worked out including the use of sensors to capture information about humans' mental, affective, and motivational states. Key research factors are (a) the human perspective that are observed in given use cases and (b) the technical perspective for modelling and testing new concepts using AI focusing on decision processes. The results are published in D 3.1 and progress of the research will be reported in D3.2.
The Software Architecture distinguishes between (a) core components and (b) a list of (AI) services the user can select to compose a solution. The (a) cores components are (i) user interface, (ii) orchestrator, (iii) knowledge base that integrates data from legacy applications and sensor data streams; (iv) external data asset marketplace complementing the data coming from inside the use case, (v) the configurator enables use-case specific solutions and (vi) the AI-enrichment service catalogue enabling a list of services. The (b) list of (AI) services is a collection of commercial products and research prototypes including 3rd party solutions. The software architecture is published in D 4.1. A first reference implementation demonstrating the operational principles behind DAI-DSS has been implemented. The demonstration is depicted in D 4.2. An update of the architecture will be published in D 4.2.
Relevant literature and developments related to (a) democratization of decision-making and digital twins for human experts, (b) technical approaches like Artificial Intelligence (AI) and Multi-Agent System (MAS), (c) aspects of reliability and trustworthiness in AI is provided. Research has been worked out including the use of sensors to capture information about humans' mental, affective, and motivational states. Key research factors are (a) the human perspective that are observed in given use cases and (b) the technical perspective for modelling and testing new concepts using AI focusing on decision processes. The results are published in D 3.1 and progress of the research will be reported in D3.2.
The Software Architecture distinguishes between (a) core components and (b) a list of (AI) services the user can select to compose a solution. The (a) cores components are (i) user interface, (ii) orchestrator, (iii) knowledge base that integrates data from legacy applications and sensor data streams; (iv) external data asset marketplace complementing the data coming from inside the use case, (v) the configurator enables use-case specific solutions and (vi) the AI-enrichment service catalogue enabling a list of services. The (b) list of (AI) services is a collection of commercial products and research prototypes including 3rd party solutions. The software architecture is published in D 4.1. A first reference implementation demonstrating the operational principles behind DAI-DSS has been implemented. The demonstration is depicted in D 4.2. An update of the architecture will be published in D 4.2.
The following results have been identified to be provided with intended TRL by the end of the project:
DAI-DSS – User Interface:
A Micro-Frontend Framework (OLIVE) to display results of AI services. Intended TRL 6-7 by BOC.
DAI-DSS – Orchestrator:
A model-based Microservice Integration Framework (OLIVE) to orchestrate data access and AI services. Intended TRL 6-7 by BOC.
Process Orchestration (Process Maestro) for coordinating and refining workflows enabling enhanced stakeholder participation. Intended TRL 6-7 by MORE.
DAI-DSS – Knowledge Base:
A Product Data Management system (EDMtruePLM) based on Open Standards for integrating, storing, and accessing product data. Intended TRL 7-8 by JOTNE.
DAI-DSS – Configurator:
Add-On Service for AI to extend process modelling (add on of ADONIS). Intended TRL 4-6 by BOC.
DAI-DSS – AI Enrichment:
Service for AI-based optimization or prediction in form of diverse AI algorithms and classical decision support techniques. Intended TRL 3-6 by RWTH.
A User Centric Service to introduce AI into legacy systems (add on of ADOIT). Intended TRL 6-7 by BOC.
Models and Concepts to integrate AI, Data, Human and Robots:
A set of optimisation algorithms using different techniques ranging from AI to mathematical optimisation – Intended TRL 6-7 – as well as IoT Innovation Space Consulting Services - Intended TRL 6-8 – by JR.
Intelligent Sensor Box for integration of stationary and wearables sensors and AI-based analytics together with consulting within Human Factor Lab. Intended TRL 6-7 by JR.
Methods and Guidelines to assess and approve AI aiming for Trustworthiness in form of tables providing transparency measures that foster trust in AI is provided. Intended TRL 4-6 by RWTH.
Physical Experimentation and Co-Creation Laboratory:
Tool support for design thinking workshop (Scene2Model) and corresponding workshop facilitation – intended TRL 5-6 – and an experimentation space to co-create and evaluate digital and AI solutions – intended TRL 4-5 – is provided by OMiLAB.
Process-Based Decision Models for multiple input parameters for internal usage with intended TRL 6 -7 is provided by FLEX and CRF.
DAI-DSS – User Interface:
A Micro-Frontend Framework (OLIVE) to display results of AI services. Intended TRL 6-7 by BOC.
DAI-DSS – Orchestrator:
A model-based Microservice Integration Framework (OLIVE) to orchestrate data access and AI services. Intended TRL 6-7 by BOC.
Process Orchestration (Process Maestro) for coordinating and refining workflows enabling enhanced stakeholder participation. Intended TRL 6-7 by MORE.
DAI-DSS – Knowledge Base:
A Product Data Management system (EDMtruePLM) based on Open Standards for integrating, storing, and accessing product data. Intended TRL 7-8 by JOTNE.
DAI-DSS – Configurator:
Add-On Service for AI to extend process modelling (add on of ADONIS). Intended TRL 4-6 by BOC.
DAI-DSS – AI Enrichment:
Service for AI-based optimization or prediction in form of diverse AI algorithms and classical decision support techniques. Intended TRL 3-6 by RWTH.
A User Centric Service to introduce AI into legacy systems (add on of ADOIT). Intended TRL 6-7 by BOC.
Models and Concepts to integrate AI, Data, Human and Robots:
A set of optimisation algorithms using different techniques ranging from AI to mathematical optimisation – Intended TRL 6-7 – as well as IoT Innovation Space Consulting Services - Intended TRL 6-8 – by JR.
Intelligent Sensor Box for integration of stationary and wearables sensors and AI-based analytics together with consulting within Human Factor Lab. Intended TRL 6-7 by JR.
Methods and Guidelines to assess and approve AI aiming for Trustworthiness in form of tables providing transparency measures that foster trust in AI is provided. Intended TRL 4-6 by RWTH.
Physical Experimentation and Co-Creation Laboratory:
Tool support for design thinking workshop (Scene2Model) and corresponding workshop facilitation – intended TRL 5-6 – and an experimentation space to co-create and evaluate digital and AI solutions – intended TRL 4-5 – is provided by OMiLAB.
Process-Based Decision Models for multiple input parameters for internal usage with intended TRL 6 -7 is provided by FLEX and CRF.