The Experiment's primary objective is to bolster the efficiency of sequencing and planning tasks in a manufacturing plant, by conceptualizing and deploying intelligent tools that can aid the production manager and inventory manager in their day-to-day operations. In order to accurately reflect the dynamic nature of the shop floor's daily activities, the integration of machine learning and optimization techniques is crucial for the development of the anticipated troubleshooting features: Demand Forecasting Tool: Leveraging advanced Machine Learning methods to assimilate market trends, annual growth rates, seasonal influences, and other pertinent factors. This tool will serve as the cornerstone for precise demand forecasting, thus facilitating supply, planning, and overall optimization processes. Sequencing Tool: Real-time optimization achieved through a fusion of Constraint Programming and metaheuristics. This approach enables efficient sequencing while considering the duality between machines and products. The primary challenge lies in implementing real-time optimization that aligns seamlessly with the practical demands of the shop floor. This entails the effective utilization of both structured and unstructured production data (including order details, operations, quantities, operation duration, and assets), alongside continuous machine data. Such an approach aims to revolutionize the management of production processes, leading to substantial advancements in overall efficiency.

The aim of the Experiment is to shorten the engineering cycle and investigate a large panel of product option and to bring services to SMEs for new products and processes design. The plan is to use AI for continuous improvement of some issues based on historical project execution and rule basis system, especially for plastic products. The challenge is how to merge all the available information: engineering information, project management information, different deliverables, technical data, machine and production data.

The aim of the Experiment is the implementation of a system able to support operators in the warehouse to identify exactly the unique sheet they need at that time for further delivery to production line. Dimensions, thickness and material need to be carefully selected following the production line’s indication; for doing so the sheets need to be handled in the most efficient way and in the shortest time.

The aim of the Experiment is the optimization of after sales activities and not on the control and implementation logic of the robot. With the help of AI, prediction models can be developed that enable rapid adaptation of the solution in different scenarios. This is also the biggest challenge, that the models adapt quickly to new application scenarios. By using the models in the field and a continuous data pipeline from the field back, for further training of the models, a continuous optimization process can be enabled.

The aim of the Experiment is the implementation of an expert system able to support operators during maintenance and technical assistance operations. The decision support system developed in this experiment aims at facilitating the identification of faults in a machine. Maintenance, repair and troubleshooting normally demand the presence of an expert. However, the growing complexity of machines and the difficulty in finding skilled operators raises serious concerns in this regard. At present, troubleshooting tools are developed as context specific solutions and are often based on rigid and inefficient logic, hindering their adoption and use. The tool developed in AI REGIO consists of a probabilistic decision support system. The tool is based on the identification of machine faults starting from the analysis of visible symptoms. To this end, the knowledge concerning the machine/process is organised as an association between each fault and the probability of questions related to specific symptoms being answered positively. Thus, for each fault, the expert knowledge is represented as the likelihood of showing specific symptoms. To further improve the proposed probabilistic troubleshooting system, AI technologies were incorporated, enabling the desired features: Speech-to-text: speech recognition to accommodate the voice-based interface improving the user experience. Natural Language Processing (NLP): to extract insights from the acquired text and select the most relevant questions to ask for the troubleshooting process. Self-learning: enhance the system performances over time by modifying, based on previous attempts, the initial probability associated with faults as well as the relation among faults and questions. Thus, the tool uses NLP to select the most appropriate questions based on a user description of the issue provided through a speech-to-text interface. Then, by means of an algorithm to minimize the number of questions necessary and hence, speed up the identification process, the system selects and asks a series of questions regarding the symptoms. Following the operator’s responses, which can be “Yes”, “Don’t Know”, or “No”, the system identifies the most probable fault.Based on the session results and feedback provided by the operator, the system update its knowledge base to enhance future performance.

The Experiment focuses on automatic intrusion detection. Today, attack detection is very simple (track cut). Depending of the use of the AT (Anti Tampering) device, some false attacks can be detected: because of temperature changes, shocks or vibrations, etc. The use of the anti-tampering cover alone is not sufficient to detect an attack, it must be combined with other sensors (position, X-ray, ...). Then many data can be collected, but there is a need of understanding the situation, to know if it’s a real attack or a “normal” use of the systems. And the last problem is to send high secure data outside the system, so that the tampering detection can be sent followed by the customer. More than using suitable AI algorithms, we could detect a real attack on the system, whatever the means used so the security will be increase.

The aim of the Experiment is to reduce existing errors allowing to increase the level of product quality, the main device is the robot arm, AI methods are applied on the device. Below are the needs to which AI technologies can be implemented: Generate the post-processing machining trajectory for individual workpiece. Use less external measurement sensors and devices to realize the tool center point (TCP) trajectory compensation. Trajectory optimization (compensation) can suit in different machining situations, which have different cutting speed, feed rate, workpiece material, workpiece geometry, machining trajectory etc. Transformation from simulation to the real hardware, the AI model is expected to keep the performance on real hardware. The challenge is to clarify in which subtasks AI can be implemented, evaluate the effectiveness of the AI methods, ensure that all AI-techniques have good compatibility to the current company platform and ensure that the entire toolchain including all AI modules is stable and robust.

The aim of the Experiment is the development of theoretical models for soil moisture estimation, those developments have occurred in laboratories considering some pre-defined constants in the model by observing the nature of scattering mechanism. The neural network can better adapt to those changes and provide better compared to the theoretical models. The data form the flow meters can be observed online by measuring daily consumption. Human operators found out that the cumulative consumption over the night can be a good indicator for increasing water consumption. Behaviour of water consumption at the leakage points is known, but it is difficult to find a theoretical model, therefore, the AI can be used to estimate those models and provide reliable prediction of water leakages.