Voice-Enabled Digital Assistant for the Manufacturing Industry
AI-focused education and training concept
Change management process
Augmented Manufacturing Analytics
A set of DIA functions interfacing the prescriptive quality analytics service connected to shop floor data sources. It will enable non-data-scientist workers to utilize and customize data analytics during product quality tests.
AI-assisted on-the-job training for new workers
A set of DIA functions interfacing the cognitive advisor service that accesses shop floor data and adapts to the learning progress of the user. It will enable machine operators and production line managers to become effective faster, which will speed up changes in manufacturing.
WHY engine prototype
The new, experimental solution component that will allow the assistant to answer “why” questions concerning advices and predictions provided by the DIA. It will increase trust in the applied AI-based functions.
Change management process for AI-collaboration
This complements the technology deployment and seeks to prepare middle-managers and other decision-makers for opportunities, challenges, and risks related to using digital assistants in the shop floor. Means to increase the acceptance of the COALA solution, avoid mistrust, and reveal new barriers concerning the use of digital assistants in industry.
Didactic concept for new worker education in AI collaboration
This is a set of instructions, competencies tests, and learning materials for education facilities to teach shop floor (blue-collar) workers competencies in managing AI opportunities, challenges, and risks. It will allow education facilities and companies to better prepare their labor force for human-AI collaboration.
DIA dialog management layer
COALA DIA Core is at the top of the solution and represents the intelligent, voice-enabled, mobile user interface.
The COALA DIA Core bases on the open source, privacy focused assistant, Mycroft, and will feature four actions: Explain (to build trust), Advice (to help users perform their tasks), Predict (to inform about future events), and Inform (to provide facts and resource status).
COALA will extend the Mycroft’s existing functions to meet manufacturing requirements, such as time criticality, reliability when dealing with factory noise, safety when giving advice to workers and security in business environment.
AI-based reasoning and planning layer
It processes the solution with Artificial Intelligence features and components, which focus on predictive quality and agile manufacturing.
is a new, experimental solution component that answers the users “why” questions about the AI’ decisions and the provided advice, prediction and information.
aims to give advice to new workers in knowledge-intensive manufacturing processes, where the production plans are short-timed and workers must quickly realise them while maintaining the product quality.
Prescriptive Quality Analytics
will enable to predict future
quality, and to prescribe mitigating actions that improve quality of products and processes and enable non-data-scientist workers to utilize and customize data analytics during product quality tests.
aggregates and generates useful context information from individual manufacturing resources, to be provided to manufacturing workers that want to make sense of a prediction or advice provided by the assistant.
Data management & integration layer
It contains heterogeneous data storages, where personal data anonymization will be handled by the data anonymization service.
Manufacturing resources layer
It contains machines and manufactured products, as well as the workers and factory environment in a wider sense, which provide data, information and knowledge to the COALA system.