Route planning is a fundamental function for autonomous vehicles (AVs) tasked with navigating complex road networks. Traditional formulations of the route planning problem typically assume a single destination, where the solution is defined as the optimal path from the starting point to the designated goal. In this paper, we introduce a two-stage hierarchical route planning algorithm designed to determine a feasible and optimal route that sequentially connects multiple target points within a road network. Our approach employs a graph-based representation of the road network and systematically integrates global and local search strategies to guarantee both the feasibility and minimality of the resulting route. The proposed approach also introduces a semantic representation of the planned route, providing natural language indications and specifying the planned behaviour of the car. This is accomplished by classifying the discrete points that define the planned route. The validity of the proposed approach was tested experimentally on a 1:10 scale autonomous vehicle.

In order to successfully integrate circular economy principles into current value chains, it is crucial to ensure the economic sustainability of disassembly processes. Recent scientific and technological innovations, including Large Language Models (LLMs) in artificial intelligence (AI), are greatly accelerating progress in enhancing robotic capabilities. Dismantling is the first step in the re-manufacturing, repair, and recycling processes of end-of-life (EoL) products. Traditionally, this operation is performed manually by operators or by expensive dedicated robotic cells. Although manual disassembly offers flexibility in handling complex situations, it is a time-consuming and labour-intensive process that can negatively affect the health of operators and the cost-effectiveness of the disassembly process. However, robot disassembly presents difficulties in handling complex parts in a flexible manner. Robot application emerges as an automated versatile solution, capable of handling uncertainties in the frequency, quantity, and quality of end-of-life products. A fully automated approach for the removal of clothing zips from garments is proposed. The approach detects the pixels corresponding to the zip using LLM-Based AI techniques to plan the path of the robotic arm. The solution reduces the effort of AI training in industrial applications. Simulations validate the approach and its performance.

Advances in machine learning (ML) have opened up new opportunities to include decision-making capabilities in Internet of Things (IoT) nodes. This opportunity is complex to address, since conventional ML implementations are computationally intensive, thus reducing the possibility of implementing them on resource-constrained systems. In addition, when nodes in a network grow significantly, centralized data processing creates new challenges such as latency, resource efficiency, privacy, bandwidth, etc. Aiming to simultaneously address the above challenges, this paper presents a novel efficient distributed learning strategy for ESN-Based model, that is conceived for training and inference to incorporate part of the model while reducing the sharing of private information. The experiment setup was conducted using Appliances Energy Prediction dataset to forecast energy consumption under different environmental and usage conditions, using simulated IoT devices. The numerical results show that the proposed solution performs well compared to a classical centralized ESN-based approach without sacrificing too much performance at very low computational costs.

The co-transport (collaborative-transport) of deformable materials such as fabrics, composite materials, cables or wires and so on, is a challenging task for robotic applications in industry. The main difficulty lies in the deformability of the material, which can slide, stretch and deform during handling and transport. Common approaches in the literature either take advantage of force sensors to act on the fabric and restore its ideal state, or estimate the deformation state of the fabric with depth images and neural networks (NN). In both cases, issues may arise regarding the effects of force control on the material and the industrial reliability of the NNs, respectively. This paper proposes a method based on the estimation of the deformability constraints of the flexible material to obtain geometric parameters that allow planning the trajectory of a collaborative manipulator for co-transport that guarantees the deformation constraints during transport. By identifying a representative point on the side of the fabric held by the man, the action necessary to restore the desired deformation of the fabric that is initially estimated is planned. Once the material assumes the desired state, the robot's movements are generated to track the human's movement and act in safety conditions. A near-time-optimal control strategy is proposed. Finally, the system is tested in a simulation scenario.

In modern industrial manufacturing, simulation and accurate forecasting of energy consumption are critical to optimise resources and reduce waste meeting the significant energy demands and greenhouse gas emissions of this sector. Industry 5.0 (15.0), focused on sustainability and human-centered manufacturing, emphasises the use of Artificial Intelligence (AI), the Industrial Internet of Things (1IoT) and the Cyber Physical Systems (CPS) to monitor and optimise resource use in real time. This paper proposes a novel energy consumption simulation method using Neural Network (NN), specifically the Deep Echo State Network (DeepESN), integrated with a text embedding model, which allows to combine energy consumption data into production scheduling. Unlike existing approaches, our method considers both machine-level energy consumption and production workflows, enabling comprehensive optimisation of energy effi-ciency. Preliminary tests in a real production scenario show the potential of the approach, highlighting its ability to predict energy consumption simply by using smart meters without additional hardware. This work represents a significant advancement in in-tegrating energy consumption modeling into production planning and contributes to more sustainable industrial practices.

The advent of increasingly flexible and adaptive production processes will require equally flexible robotic collaboration. The forthcoming robotic applications must be able to self-adapt to different applications and scenarios and to handle frequent interaction with humans and dynamic environments. Environment perception, object recognition, and trajectory re-planning in a dynamic and changing environment, even with possible human interaction, are features that are not yet all simultaneously available in collaborative robots, and certainly not in industrial robots. This paper proposes some initial results of the authors’ ongoing research on the development of a ROS2-based framework for industrial applications due to which a robotic manipulator equipped with a depth camera is able to have simultaneously perception, recognition and re-planning capabilities in a dynamic and changing environment. This framework is potentially applicable to various industrial robots and depth cameras; here, the first results of the experimental implementation on an Omron TM5-900 collaborative robot equipped with a fixed depth camera will be shown.

The desired path following for an autonomous vehicle can be achieved by ensuring the desired yaw rate response. In conventional methods the control of lateral vehicle dynamics by means of yaw rate-based PID controller has sometimes been investigated but rarely implemented. An autonomous driving system can therefore be achieved by designing a reliable yaw rate controller for the vehicle. In this paper, an autonomous vehicle control architecture and a steering angle PID controller based on the vehicle’s yaw rate measured by a gyroscope is proposed and designed to handle most of the manoeuvres of an autonomous vehicle. It does not require measurements of lateral acceleration and lateral speed. The yaw rate reference is provided as an output by a higher-level system and used in the lower-level control loop. The validity of the proposed approach was tested experimentally on a 1:10 scale autonomous vehicle.

This paper focuses on the study and development of learning algorithms oriented to wear classification and predictive maintenance (PdM) of the cutting tool (CT) of a clamping machine for producing structural steel bars. While several works dedicated to CTs for turning and milling operations, or in general for metal removal operations, also known as subtractive manufacturing processes, can be found in the literature, the phenomena related to cutting with a cutting knife have not been widely treated in the literature. This article intends to focus on the analysis of the latter problem. The objective is to estimate the wear of the CT, a critical component of the steel bar cutting machine. The SVM classifiers were therefore used to classify the wear. For the predictive maintenance purpose, two algorithms were implemented for the prediction of the remaining service life, based on the Degradation Model and the Similarity Model respectively; in the first method, a prediction and state update function were used, while in the second method, a Long Short-Term Memory (LSTM) Neural Network (NN) was used.

The present paper proposes an architecture for a product process monitoring system suitable for SMEs (Small-Medium Enterprises). The monitoring system is the main means by which decision-making systems based on intelligent automation technologies are aware of the state of the system on which they will take decisions. Methods and tools from best-practice and best-effort approaches are proposed in the context of SMEs, where the requirements of low cost, low initial level of digitalization and high production flexibility often coexist and contribute to the complexity of management and control problems in these companies. The paper focuses on the design of the monitoring system using an OSGi framework to meet industry standards and Industry 4.0 requirements, taking into account the peculiarities of SMEs as design constraints. The proposed architecture was first tested using a simulation tool and then implemented on a full-scale production line used for data collection.