With the help of Croatian AI, the business of the future will be led

—

FSB ESARP

foto

—The most resilient part of the global economy consists of manufacturing companies. They are the backbone of any healthy economy, but to survive in the global market, they must face numerous challenges. Among the biggest are achieving high efficiency in production and operations, maintaining competitiveness, and constant adaptation to innovations. Innovations are precisely what can best empower manufacturing companies to keep pace with the relentless competition in the market.

Today, it is impossible to look towards the business future without artificial intelligence (AI), which is already being used at many levels of manufacturing companies’ operations. However, there is no artificial intelligence that can unify all business processes, make interdependent decisions across each process, and then offer solutions for complex operational and strategic problems. Nevertheless, such a solution may soon exist, and it will be taken care of by experts from the Institute of Industrial Engineering at the Faculty of Mechanical Engineering and Naval Architecture (FSB) in Zagreb.

Explainable Artificial Intelligence

The research team of Assistant Professor Miro Hegedić, which includes Dr. Mihael Gudlin, a computer vision expert, and researchers Matija Golec and Petar Gregurić, who is currently conducting research related to his doctoral thesis on this topic, is most involved in the development of this product and leads the resulting spin-off company. Last year, they began the initial development of the ESARP project, an expert system for autonomous management of business processes in manufacturing companies. This solution is based on the development of explainable artificial intelligence (XAI) that can clearly explain its decisions, conclusions, or recommendations and enable management and operational decision-making in seven selected business functions of manufacturing companies: research and development, marketing and sales, human resources, procurement, production, logistics, and finance.

With the help of ESARP, various scenarios will be predicted and decisions will be made regarding the management of business processes, which should reduce the risk of subjectivity, inexperience, overlooking the bigger picture, or ignorance of decision-makers. Given that it is easier to operate in the market as a company rather than an institution, scientists have launched the aforementioned spin-off startup company, one of the first of its kind in the Croatian market. The potential was early recognized and supported by Impuls Consulting Ltd., specialized in financing innovation projects.

Participation in Competitions

The project has also been submitted for obtaining non-repayable funds according to the call from the Ministry of Science and Education for ‘Start-up/spin-off companies of young researchers’, worth a total of 5,043,466 euros. The competition stipulates that an individual project can receive at least 66,000 euros and a maximum of 200,000 euros, and such a financial injection can significantly impact the increase of knowledge transfer from scientific institutions to the private sector. Such competitions and financial incentives allow young researchers to turn their ideas into reality and thus influence the further development of the Croatian economy.

– We are currently waiting for the results of the competition for research and development support, which will enable us to develop the first fully autonomous so-called ‘water-spider’. This is a collaborative robot used for manual operations when workers are overloaded, taking on part of the burden. However, we will be able to say more about this if we are granted funding and when the system is fully developed – states Assistant Professor Hegedić.

Darko Liović, CEO of Impuls Consulting, emphasizes that he has been collaborating with Miro Hegedić’s team for years and is very satisfied with the results achieved so far. Many clients of Impuls Consulting have worked with FSB, he added, and everyone wants to work together again.

– When Petar Gregurić and Miro Hegedić approached us with a ‘pitch’ for an autonomous management system for manufacturing companies, I immediately recognized the potential. Primarily from a business perspective, as I believe it can be a commercially viable venture, but also because I am enrolled in a doctoral program at FSB and am dealing with this area. We helped them with the application for the spin-off competition and eagerly await the results – says Liović.

Managerial Difficulties

The project ‘Development of ESARP – an expert system for automated management of processes’ is planned to be implemented over 24 months with the aim of conducting research and development activities to develop a system intended for new generation companies primarily engaged in line and serial production. Such productions are now largely automated and robotized with a small share of human labor in repetitive tasks.

The expert system will make business decisions while simultaneously considering all business functions in the company, taking into account the interdependence of all business decisions in the company and selecting the best business choices at any given moment.

In this way, key performance indicators (KPIs) set by the company owner will be achieved, which do not necessarily have to be related to profit increase but can be set differently. For example, increasing market share, achieving the lowest product price in the market, achieving full employment, or sufficient market supply. Predictive analytics will also be applied in the business decisions of the manufacturing company by analyzing historical data such as market price movements in procurement and sales, predicting consumer habits and preferences in the market, and trends. For successful business operations, experts from FSB are convinced that it is essential to utilize the capabilities of machine learning and artificial intelligence in interaction with humans.

—

Miro Hegedić

foto

—

Furthermore, according to surveys and interviews with managers, the most common problems faced by the metal and food industries, as the two largest branches of the manufacturing industry, have been identified. In the metal industry, managers face high raw material costs, complex metal processing processes, the need for specialized equipment, and high-quality standards, along with challenges in storing and managing large quantities of metal, recycling waste, and maintaining workplace safety. In the food industry, the biggest challenges are short product shelf lives, complex management of various ingredients and recipes, compliance with food safety regulations and high-quality standards, inventory and warehouse management, monitoring and controlling product correctness, and maintaining hygiene standards. The demand for ESARP is evidenced by the fact that several manufacturing and food companies in Croatia have sent letters of intent expressing interest in ESARP.

Younger Workforce

The project also reveals research problems from a technical perspective that are complex and require a multidisciplinary approach as it combines knowledge from the fields of machine learning, computer science, business processes, and social sciences to achieve progress in the development of explainable artificial intelligence (XAI) through the development of an expert system. This is a challenge of AI interpretability as the decisions made must be explained understandably, integrating AI into business processes, and scalability and efficiency since large amounts of data and complex processes are associated with manufacturing companies.

ESARP will enable employers to place younger managers with less management experience in necessary positions and provide the younger workforce with the opportunity for faster advancement and acquisition of new skills. The acquisition of new skills and knowledge will be facilitated by the explanations and understanding of the decisions made by the expert system, and in the long term, the assumptions and decisions generated by XAI will be able to be integrated into blockchain, ensuring state-of-the-art security for decision-making traceability and secure access to historical data.

Research and Development Projects

The Institute collaborates with some leading Croatian manufacturing and food companies and has conducted several large research and development projects. One project, worth over 50 million kuna, for example, was related to the introduction of lean manufacturing, optimization of planning, and the development of innovative solutions for breaking metal chips and designing a new way of measuring the quality of workpieces with lasers with very high tolerances required for the automotive industry.

In another project with Končar – Energy Transformers and Windor, an experimental factory for learning was created, where production processes can be tested before being put into production. The factory integrates all known technologies of Industry 4.0, including the 5G campus network, and can therefore also be used for lifelong learning. The VeryQ system, which is currently in the development process, is an integrated system that allows the creation of computer vision algorithms even for individuals without programming experience, as it uses a no-code approach. The goal is for a manufacturing company, regardless of the product, to be able to apply VeryQ in one day, assemble the first version of the algorithm, and put it into production.

—

Petar Gregurić

foto

—

– At the end of 2019, a project for the development of a modular expert system for establishing discrete manufacturing processes based on the application of Smart Factory principles began at FSB. Modern equipment from the field of ‘Smart Factory’ has been installed at the Institute of Industrial Engineering, which, after connecting to a private communication network, will enable students, postgraduates, and entrepreneurs to develop new knowledge and experiences. They will later apply these at their workplaces and thus modernize the industry of the Republic of Croatia through digital transformation – said Professor Dr. Nedeljko Štefanić, head of the Smart Factory project, who has led several research and development projects at the Institute of Industrial Engineering.

FSB has also included collaborative robots in the third development project in the smart factory and has constructed future work assembly stations, in accordance with the principles of Industry 5.0, where everything is connected to humans. It has been recognized that a significant problem in production is the disconnection of systems and decision-making based more on experience than on numbers.

*Content created in collaboration with the Faculty of Mechanical Engineering and Naval Architecture in Zagreb