Information about participation in the 2018-1.1.2-KFI-2018-00134 application
Beneficiary name: Training360 Kft. (Consortium partner: e-Corvina Informatikai Szolgáltató Kft.)
Project title:Development of a simulation training methodology and framework based on industrial IoT and machine learning
Contracted amount of support: 481 543 350 HUF
Grant rate: 66,71%
Project completion date: 31.12.2020.
Description of the project
Training360 Ltd., together with e-Corvina Informatikai Szolgáltató Ltd., has been awarded a grant of more than HUF 481.5 million under the call for proposals "Support for RDI activities of small, medium and large enterprises". The project, with a total budget of nearly HUF 721.78 million, will develop a new methodology and framework.
Industry 4.0 is no longer just a slogan: we are now at a stage of development where a new type of integration of IT and automation is enabling the use of artificial intelligence systems. This evolution could lead to solutions for controlling, optimising and fine-tuning processes without human intervention. However, further progress is needed in a number of areas, ranging from autonomous robotic systems to systems integration. This development vision links three areas that are crucial to the creation of Industry 4.0: data collection, data analysis with artificial intelligence, and decision support.
The project involves the automatic collection of data generated by industrial sensors (Industrial IoT) and human labour on production lines, followed by quantitative and qualitative analysis of the data and the use of machine learning-based artificial intelligence to link and optimise the incoming data and the associated decision cases. To better understand and improve decision support, we will create a case study-based simulation training framework to further test and analyse the collected decision scenarios. In this way, we propose a solution not only for optimising decision support but also for training human resources.
The successful application of machine learning algorithms requires the proper selection of the algorithms used, the creation of specific learning sets, the choice of input and output parameters. This is a complex task, as the data generated in production must be directly linked to management-level decisions and these links must be tested.
The project aims to address these challenges by developing a simulation training methodology that uses both sensor data and human feedback in an arbitrary industrial environment. The developed multi-actor simulation scenarios will be tested live by the trainees, feedback and decisions will be recorded, and conclusions will be drawn and analysed. In this way, a growing knowledge base is created, which includes the scenarios and management decisions, together with the sensory and human data from the underlying production floor. The resulting knowledge base can also be used to develop bots (programmed actors) that can interact with other actors in different roles, so that simulations can be extended to production line workers, middle managers and senior managers. This will allow to develop appropriate responses and have a real impact on the quality of decision-making. The same knowledge base can be used to train machine learning algorithms.
In the course of the project, data collection, development of machine learning algorithms and simulation scenarios will be carried out in a specific industrial plant. This production plant will provide the sensor and decision input and output data. On this sample, we will test and develop a general methodology that can be used regardless of industry. The necessary IT infrastructure will also be developed in conjunction with the industrial research and pilot development phases. In the pilot development phase of the project, the practical applicability of the scenarios created will be tested with the employees of the organisation, recording responses and decisions at different organisational levels.
The project will result in a new methodology and framework that will allow the application of the created data collection, machine learning and simulation training solution in different industries, linking industrial IoT, machine learning, employee training and decision making.