ASSISTANT
ASSISTANT's main objective is to develop a collection of intelligent digital twins that will automatically adapt to the manufacturing environment. These digital twins will assist in the design and operation of a complex, collaborative, reconfigurable, mixed/multi-model production system based on data collected by IoT devices.
ASSISTANT promotes a methodology that will enhance generative design with Artificial Intelligence learning aspects from data available in manufacturing.
Méthode utilisée
The methodologies in ASSISTANT are based on the joint use of machine learning (ML), optimization, simulation, and domain models. The ASSISTANT approach wil go through data collection, followed by enrichment of the digital twin and culminating in varios AI-based decision support tools for process design, production planning/scheduling and real-time control in manufacturing systems. Thes decision support tools (demonstrators) will be tested based on (anonymized) data from "real" business environments in ASSISTANT as autonomous prototypes. After succesful validation of the demonstrators, the use cases, system/process specification, system architecture, concepts, methods, techniques, algorithms and software code will be available to the ASSISTANT consortium partners. They can be incorporated into the manufacturing systems studied for application in daily operations by the consortium partners at their own expense. Most of the results of the project will be disseminated as widely as possible in the industry to increase the number of users
IMT Atlantique Role
IMT Atlantique is the project leader and therefore the coordinating institution of the ASSISTANT project.
Although also involved in the technical solutions of the project, IMT Atlantique ensures the achievement of the project objectives in terms of scientific quality, timely delivery, and contribution to the expected impact of the project. IMT Atlantique ensures an efficient follow-up of the project's progress. The management of the consortium and the scientific and technical coordination are under its responsibility.
Expected outcomes
In a synthetic way, ASSISTANT aims to provide the results listed on the following 5 points:
- An optimization software that is synthesized. Currently, most production planning models are developed manually for each type of production environment based on certain fixed assumptions. With the qualitative variability of demand, this is no longer feasible, ASSISTANT aims to synthesize the production of models and software components directly from the acquired data.
- Digital twins that will rely on process planning, production planning and scheduling and finally, real-time control of the manufacturing plant to consider the entire production line
- Explanation-based interaction with humans. Digital twins and machine learning should be able to bring humans into the loop by providing concise human-interpretable explanations, allowing humans to resolve some ambiguities
- A multi-level distributed architecture that extends from the specification of the manufacturing/assembly system to the execution of a production plan
- Validation and application of the tools on three concrete industrial cases while providing a toolbox of components that can be used in other possible use cases outside the project.