The mobile manipulator’s base positioning is strategic to optimize the execution of different tasks within a specific workspace. Base placements provided by the human intuition yield suboptimal results, and existing methods are usually based on computationally expensive maps built to represent the distribution of the manipulator’s dexterity or the base positions effective for reaching a specific target, but not both simultaneously. This paper proposes a novel optimization-based method for mobile manipulator’s base placement, focused on a compact mathematical model of the manipulator reachability space. By representing it through two concentric ellipsoid equations, such a model enables efficient evaluation of end-effector reachability while guiding the base placement. This model is exploited in the optimization problem to autonomously reposition the mobile base, ensuring pose reachability with high dexterity and minimizing collision risk. The approach is validated experimentally using a mobile manipulator in a real-world laboratory and assigning target poses with varying difficulty levels. The results demonstrate the effectiveness of the obtained base poses and the adaptability of the method to different scenarios.

Nowadays, mobile manipulators can support humans in daily life tasks while sharing the same workspace. These robots are usually requested to perform pick-and-place actions that involve objects that must be handled with care, since they may hurt the human operators. To optimize their utility as smart assistants, they require autonomous grasping pose generation, object recognition, and pose estimation capabilities. In addition, since they work in dynamic environments, adaptability is essential, hence predefined starting positions for grasping actions should be avoided. These demands are even more challenging for robots with limited computational capabilities. In this paper, we propose an approach that demonstrates how to improve the capabilities of a low-resource mobile manipulator. First, an easy way to model the robot as a unique system for holistic motion planning is developed. Then, we propose a lightweight approach to generate the grasping point to pick a requested item that relies only on the available CPU. Finally, a simple yet flexible solution that involves human feedback is adopted to let the robot handle potentially dangerous objects, while ensuring the operator's safety. The proposed solution has been developed in ROS1 and experimentally tested on the LoCoBot mobile manipulator in a laboratory environment.

Object manipulation in unstructured environments is important for many industrial applications where the items vary in shape, size, and material. This paper introduces a two-step pipeline for object picking, which combines instance segmentation with a heuristic based grasp point selection. The grasping points are determined using the 2D segmentation masks and depth images. A voxel-downsampling procedure reduces the depth noise, and the Theil-Sen algorithm ensures a robust linear regression for the grasping attitude determination. Unlike other methods, our approach does not require extensive training, as well as a fine labelled dataset for picking, and hence it is also independent of object shapes. Using SAM's ViT-h version and a binary object detector trained on a large dataset, our method is robust and class agnostic. The experiments, made using a RealSense D435i camera and a Racer 3 manipulator, show that our pipeline has a good success rate in simple and moderately complex scenarios, balancing computational efficiency and accu-racy.

Smooth and collision-free trajectory planning is crucial to high speed and high precision machining, such as 3D laser cutting. However, it is difficult to further enhance the kinematic performance of the primary translational axes during the process. This paper presents a novel two-phase planning strategy, which optimizes the tool orientation and leverages a redundant standoff axis to significantly enhance the smoothness of the translational movements in redundant 3D laser cutting machines. In the first phase, collision-free configuration spaces (C-spaces) are constructed along the tool path, utilizing a graph-based search approach with Dijkstra's algorithm for tool orientation optimization. Subsequently, a secondary orientation curve, namely the M path, is planned in the second phase with a variable distance from the primary tool path curve, and the motion of the redundant standoff axis is handled via a deep reinforcement learning approach. The proposed methodology provides an advancement in conventional five-axis machines lacking of flexibility. Experimental validation confirms the potential of the approach to substantially improve machining accuracy and efficiency.

A crucial aspect of human-robot collaboration involves the robot’s ability to safely perform handovers of objects to be used by the operator. In this study, we introduce two complementary frameworks designed to execute every stage of a robot-to-human handover process, using only RGB-D input data. The first framework employs a machine learning model, trained on a custom real-world dataset, to detect objects and their parts’ affordances. Affordance is encoded using a novel representation based on keypoints, which are utilized to model hazardous sections of objects and plan appropriate grasps. The second framework tracks hand movements to dynamically determine handover locations, while enforcing safety protocols to prevent exposure of dangerous object parts during the movement. Additionally, it ensures correct object orientation upon delivery, presenting the object handle to the human. The effectiveness of the proposed solution has been successfully validated through testing on a real mobile manipulator.

Mobile manipulators can significantly contribute to enhance the flexibility of several processes, such as automated order-picking systems and various logistic applications, thanks to their capability to manipulate objects and deliver them to different locations. A primary role is envisaged for them in Smart Factories, as workmates of the operators, if they are able to safely navigate in human-shared environments. This paper proposes a lightweight and flexible ROS1-based software architecture, designed for low-resource mobile manipulators, to make them able to autonomously search for the items requested by a human operator, independently from the starting pose, and pick and place them in a predefined depot location. The validity of the proposed architecture, which is potentially applicable to different low-resource mobile manipulators, is proven through its experimental implementation on a Locobot mobile robot.

The up-and-coming concept of Industry 5.0 fore-sees human-centric flexible production lines, where collaborative robots support human workforce. In order to allow a seamless collaboration between intelligent robots and human workers, designing solutions for non-expert users is crucial. Learning from demonstration emerged as the enabling approach to address such a problem. However, more focus should be put on finding safe solutions which optimize the cost associated with the demonstrations collection process. This paper introduces a preliminary outline of a system, namely EValueAction (EVA), designed to assist the human in the process of collecting interactive demonstrations taking advantage of simulation to safely avoid failures. A policy is pre-trained with human-demonstrations and, where needed, new informative data are interactively gathered and aggregated to iteratively improve the initial policy. A trial case study further reinforces the relevance of the work by demonstrating the crucial role of informative demonstrations for generalization.

Mobile robots can highly contribute to achieve the production flexibility envisaged by the Industry 4.0 paradigm, provided that they show an adequate level of autonomy to operate in a typical industrial environment, in which the presence of both static and dynamic obstacles must be managed. Robot Operating System (ROS) is a well known open-source platform for the development of robotic applications, recently updated to the enhanced ROS2 version, including a navigation stack (Nav2) providing most, but not all the capabilities required to a mobile robot operating in an industrial environment. In particular, it does not embed a strategy for dynamic obstacle handling. Aim of this paper is to enhance Nav2 through the development of a Dynamic Obstacle Layer, as a plug and play solution suitable for the integration of the dynamic obstacle information acquired by a generic 2D LiDAR sensor. The effectiveness of the proposed solution is validated through a campaign of simulation tests, carried out in Webots for a TurtleBot3 burger robot, equipped with a RPLIDAR A3 LiDAR sensor.

The brand new paradigm of Industry 5.0 envisages an increased leading role of the human operator in the production lines of the next future. Human-centric oriented solutions are going to be developed based on proactive human-robot collaborations, able to better exploit the skills and capabilities of both humans and cobots, mainly thanks to artificial intelligence. Several functionalities must be assured to reach such a goal, guaranteeing safety and flexibility, from human action prediction to object recognition and affordance. This paper offers an overview of the existing solutions for the various, separate issues, proposing a general framework for mobile manipulators assisting human workers, in a context of mass customization.

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.

Human-robot collaborative applications are generally based on some kind of co-working of the human operator and the robot in the execution of a given task. A disruptive change in the collaborative modalities would be given by the capability of the robot to anticipate how it could be of help for the operator. In case of an Autonomous Mobile Robot (AMR), this would imply not only a safe navigation in presence of a human operator, but the automatic adaptation of its motion to the specific operation carried out by the operator. This paper investigates the possibility of achieving operation recognition by monitoring the human motion on a 2D map and classifying his/her path on the map, taken as an image data sample. Deep learning state-of-the-art libraries and architectures are exploited with the aim of making the robotic system aware of the ongoing process. The reported results, relative to a small training dataset, are nonetheless promising.

In the last decades, industrial manipulators have been used to speed up the production process and also to perform tasks that may put humans at risk. Typical interfaces employed to teleoperate the robot are not so intuitive to use. In fact, it takes longer to learn and properly control a robot whose interface is not easy to use, and it may also increase the operator's stress and mental workload. In this paper, a touchscreen interface for supervised assembly tasks is proposed, using an LCD screen and a hand-tracking sensor. The aim is to provide an intuitive remote controlled system that enables a flexible execution of assembly tasks: high level decisions are entrusted to the human operator while the robot executes pick-and-place operations. A demonstrative industrial case study showcases the system potentiality: it was first tested in simulation, and then experimentally validated using a real robot, in a laboratory environment.

The manufacturing assembly lines of the future are foreseen to dismiss fully unmanned systems in favour of anthropocentric solutions. However, bringing in the human complexity leads to modeling and control questions that only data can answer. Moreover, many human-robot collaborative applications in flexible manufacturing involve manipulator cobots, whereas little attention is given to the role of mobile robots. This work outlines a data-driven framework, which is the core of a brand new project to be fully developed in the very next future, to let human-robot collaborative processes overcome the barriers to successful interaction, leveraging mobile and fixed-base robots.

PMSMs are widely used in applications on electric vehicles, robotics and mechatronic systems of industrial machinery. Thus it becomes increasingly interesting to prevent their fault or malfunctioning with Predictive Maintenance (PdM). However, reaching this outcome could be difficult, especially if the stationary condition is not achieved and without additional sensors. This paper examines the use of a load torque observer based on Extended Kalman Filter for the diagnosis of electric drives working under non-stationary conditions. The proposed Motor Torque Analysis (MTA) is compared with the Motor Current Signature Analysis by evaluating their diagnostic capabilities under the assumed conditions. Finally, the results of bearing failure detection under non-stationary conditions are presented, highlighting the superior diagnostic capabilities of the MTA under such conditions.

In the near future, an increasing number of mobile agents working closely with human operators is envisaged in smart factories. In industrial human-shared environments that employ traditional Automated Guided Vehicles, safety can be ensured thanks to the support provided by Autonomous Mobile Robots, acting as a net of meta-sensors. The localization and perception information of each meta-sensor is shared among all mobile platforms. In particular, the information about the dynamic detection of human presence is combined and uploaded in a shared map, increasing the awareness of the mobile robots about their surroundings in a specific working area. This paper proposes an architecture that integrates the meta-sensors with an existing net of Automated Guided Vehicles, with the aim of enhancing systems based on outdated mobile agents that seek for Industry 4.0 solutions without the necessity of a complete renewal. Simulations of test scenarios are provided in order to confirm the validity of the proposed architecture model.

In smart factories, the performance of the production lines is improved thanks to the wide application of mobile robots. In workspaces where human operators and mobile robots coexist, safety is a fundamental factor to be considered. In this context, the motion planning of Autonomous Mobile Robots is a challenging task, since it must take into account the human factor. In this paper, an implementation of a three-level online path planning is proposed, in which a set of waypoints belonging to a safe path is computed by a supervisory planner. Depending on the nature of the detected obstacles during the robot motion, the re-computation of the safe path may be enabled, after the collision avoidance action provided by the local planner is initiated. Particular attention is devoted to the detection and avoidance of human operators. The supervisory planner is triggered as the detected human gets sufficiently close to the mobile robot, allowing it to follow a new safe virtual path while conservatively circumnavigating the operator. The proposed algorithm has been experimentally validated in a laboratory environment emulating industrial scenarios.

This paper proposes the basic concepts of a brand new HW/SW architecture, whose development is in progress through an academic/industrial collaboration, aimed at obtaining a mobile agent capable to merge in itself the standard characteristics of the Automated Guided Vehicles and some potentialities of the Autonomous Mobile Robots, with a particular care for safety issues. Its HW/SW features, together with its mechanical characteristics, make it potentially applicable both in industrial and research contexts.

The presence of mobile agents in the industrial environment is growing, introducing specific safety issues in their path planning. This paper proposes the implementation of a three-level path planning procedure, which allows: (i) the imposition of a set of waypoints, tending to a safe path, generated by a supervisory planner on the basis of a static map of the environment (not necessarily fully updated), (ii) the generation of a global path including such waypoints exploiting a cost-based algorithm, taking into account also the obstacles not included in the static map, but detected at the beginning of the global planning phase, and (iii) the avoidance of dynamic obstacles appearing during the robot motion, thanks to the action of a local planner. The procedure has been experimentally tested to plan the motion of a differential mobile robot.

A growing presence of mobile agents is envisaged in the smart factories scenario of the next future. The safe motion of traditional Automated Guided Vehicles in human-shared workspaces can be achieved thanks to the support of a fleet of Autonomous Mobile Robots, acting as a net of meta-sensors, able to detect the human presence and share the information. This paper proposes a preliminary working implementation of one meta-sensor module, exploiting the synergistic use of different sensors through an overall affordable and accessible sensor data fusion algorithm. Experimental results in a laboratory environment confirm the validity of the approach.

Fast prototyping, in combination with Open Source resources, has a key role in the present technological evolution. In this work, an Open Source ROS-based architecture for Fast Software Prototyping for robot arms is proposed, based on Docker containers as a porting tool. This new architecture, called P&P (Plug and Prototype), has been integrated with an Open Source 3D-printed robot arm, to validate its ease of integration and use, also for less expert users. The P&P design is introduced and proposed as a standard for robot arms applications development, thanks to its modular and easily extensible structure.

Although an industrial robot represents a higher class of machine, it is a part of a production system. The robot, regardless of its own complexity, can be represented by a set of actions that leads itself towards the achievement of production goals. In the present paper the robot actions will be modeled as a production system in order to manage and monitoring its actions toward a production target. This will allow an easier integration of the robot within the production process, as required by the new paradigms for the factories of the future.

Programming complex robotic tasks can be very difficult, and is usually carried out by programmers very skilled in different fields, since they have to manage different issues at the same time, e.g., the process design, the task sequencing, and multi-robot programming issues like the collision avoidance and the optimization of the process according to given criteria. A task-oriented programming approach, leaving the programmer design only the process itself and the robotic cell by means of some CAD drawer softwares, makes the programming process simpler and invariant to the skill level of the programmer. The usage of complex models, like the High Level Model previously proposed by the authors, including the notions of task and world at the same time, allows to take into account both the process constraints and the physical ones. Such a model can be exploited to obtain all the feasible work-flows carrying out the required process. The formal description of the proposed work-flow model is provided in this paper, as well as the procedure to build it starting from a given High Level Model. The developed workflow model is suitable for a possible conversion into Petri Nets, as well as for the inclusion in production efficiency frameworks aimed at optimizing the overall process and avoiding possible bottlenecks. A case study is also proposed in order to show how the work-flow model can be built in a realistic, simple but not trivial scenario.

The twelve papers in this special section focus on the development of robotic solutions for smart factories in industry - the concept of the fourth industrial revolution (industry 4.0). The inclusion of robotics is expected to deeply change the future manufacturing and production processes, and lead to smart factories that will benefit from the main design principles of Industry 4.0: interoperability, virtualization, decentralization, real-time capability, service orientation, and modularity. Robotics will have a key role in this development since innovative technologies and solutions, traditionally associated with the service robotics sector, are going to migrate to industrial smarter robots, exploiting the maturing of navigation, localization, sensing, and motion control technologies. These smarter robots will draw on a much broader range of technology, allowing higher levels of dexterity and flexibility, the ability to learn tasks without formal programming, and to autonomously collaborate with other autonomous devices and human operators.

A joint research between Politecnico di Torino and COMAU is going on, aimed at defining a task oriented programming approach to allow soft skilled programmers to develop optimized programs for complex robotic cells. Within this paradigm, the user is simply asked to specify the features of the robotic cell and a structured set of tasks defining the required process. The paper is focused on the first step necessary to achieve the overall goal, i.e., the definition of a generic task model. The model is built starting from the analysis of the most important industrial applications, and can include physical and synchronization constraints, as well as various possible requirements and situations relative to the characteristics of the robots in the cell and of the involved work-pieces. A welding process is considered as typical industrial case study, to show how to build the corresponding task model according to the proposed approach.

Robotics is a surprisingly old discipline, and robots have shaped industry and the various industrial revolutions for many decades. This paper covers topics relevant to the IES Technical Committee on Factory Automation, focusing in particular on the evolution of industrial robotics. After providing a historical perspective on the topic, the paper addresses current and future trends, revealing the close link between the progress in industrial robotics and the parallel evolution of industrial communication systems, which represent an enabling technology for modern industrial robotics.

The paper proposes an off-line robot motion planning approach, aimed at the reduction of the energy consumption, for any trajectory defined on a set of target points including some “fly” ones, i.e., points that must not be exactly reached by the robot. The proposed solution is based on the search for alternative paths close to the fly target points originally defined by the operator, and on the computation of the mechanical and total energy associated to each single possible motion defined for the generated alternative paths. A branch and bound algorithm is employed to scan all the possible motions and find the complete trajectory that corresponds to the minimum energy consumption. The approach has been implemented in a software architecture that exploits some COMAU software modules, but it could be easily adapted to generic, similar modules. The results that have been experimentally obtained for two typical industrial cycles, performed by a COMAU Racer robot, can be considered as very satisfying.

The paper proposes a general friction identification framework for industrial manipulators, including the automatic handling of all the required phases, from data acquisition and processing up to parameters identification. A complete static friction model is used, with the insertion of a rough approximation of the hysteretic behavior of friction; switching to a possible simpler model is also automatically executed when possible. The proposed solution has been implemented in a software module, which has been integrated into the control architecture of an industrial robot, and experimentally tested. The results have shown that a very accurate reconstruction of the actual motor currents is provided, when the friction estimated using the proposed framework is inserted in the robot dynamic model.

The paper illustrates the research activity of a joint research laboratory, established by COMAU and Politecnico di Torino, named “RoboLAB”, and devoted to industrial robotics issues. The lab facilitates a stable cooperation between the academic and industrial partners, leading to satisfying results both from the scientific and the technological point of view. The paper illustrates in particular the most relevant results of the research activity carried out to make the industrial robots safe machines for themselves and for people working with them, dealing with collision detection and avoidance, and with solutions allowing robots to safely share spaces with humans.

A general procedure for collision detection between an industrial robot and the environment is proposed in this paper. The procedure does not use any external sensor, and does not rely on any particular information about the specific robot on which it is applied, so that it can be easily implemented in the software architecture of different manipulators without any customization. Experimental results on both lightweight and heavyweight industrial manipulators confirm its validity and the absence of false collision detections during standard work-cycles.

Robot interaction with the surrounding environment is an important and newsworthy problem in the context of industrial and service robotics. Collision avoidance gives the robot the ability to avoid contacts with objects around it, but most of the industrial controls implementing collision avoidance checks only the robot Tool Center Point (TCP) over the objects in the cell, without taking into account the shape of the tool, mounted on the robot flange. In this paper a novel approach is proposed, based on an accurate 3D simulation of the robotic cell. A distributed real time computing approach has been chosen to avoid any overloading of the robot controller. The simulator and the client application are implemented in a personal computer, connected via a TCP-IP socket to the robot controller, which hosts and manages the anti-collision policies, based on a proper speed override control. The real time effectiveness of the proposed approach has been confirmed by experimental tests, carried out for a real industrial setup in two different scenarios.

The paper is focused on the development of an adequate model of the friction acting on the joints of an industrial manipulator, suitable to be used for simulation and control purposes. The experimental tests required for the identification are executable via the standard interface for the robot programming, without any change in the path planning procedure or in the robot control. The proposed friction model is developed and validated for the six-dof Comau SMART NS12 manipulator.

Robotics education at M. Sc. courses is quite challenging, especially in the organization of intriguing laboratory activities about industrial manipulators, to accompany and complete traditional material and lessons on kinematics, dynamics, motion planning and control issues. Time and space constraints, the number of students attending the labs, and the available software and equipments (and their cost) can seriously limit the lab practices that can be actually offered in a Robotics course. On the basis of the experience matured at Politecnico di Torino, this paper describes how lab practices about manipulators can be organized by using low-cost structures, built up from the standard LEGO Mindstorms Kit, so to let the students become familiar with manipulators in different ways. Such practices are then well completed by activities with an industrial robot, available at Politecnico.

Robots and automated systems can be employed in the logistic field to efficiently perform common tasks like building and updating maps of indoor and outdoor logistic spaces, locating specific goods on the map, tracing the product flow in the area, while assuring the surveillance of the environment. This paper reports and discusses the already achieved results of the on-going research project MACP4Log (Mobile Autonomous and Cooperating robotic Platforms for supervision and monitoring of large LOGistic surfaces), aimed at the study and development of a prototype of a mobile robotic platform, with on-board vision systems and sensors, integrating a flexible wireless communication solution, able to move autonomously in large logistic spaces, and to communicate with a supervisor and other similar platforms in order to achieve a coordinated action to carry out specific tasks.

Multi robot systems are envisioned to play an important role in many robotic applications. A main prerequisite for a team deployed in a wide unknown area is the capability of autonomously navigate, exploiting the information acquired through the on-line estimation of both robot poses and surrounding environment model, according to Simultaneous Localization And Mapping (SLAM) framework. As team coordination is improved, distributed techniques for filtering are required in order to enhance autonomous exploration and large scale SLAM increasing both efficiency and robustness of operation. Although Rao-Blackwellized Particle Filters (RBPF) have been demonstrated to be an effective solution to the problem of single robot SLAM, few extensions to teams of robots exist, and these approaches are characterized by strict assumptions on both communication bandwidth and prior knowledge on relative poses of the teammates. In the present paper we address the problem of multi robot SLAM in the case of limited communication and unknown relative initial poses. Starting from the well established single robot RBPF-SLAM, we propose a simple technique which jointly estimates SLAM posterior of the robots by fusing the prioceptive and the eteroceptive information acquired by each teammate. The approach intrinsically reduces the amount of data to be exchanged among the robots, while taking into account the uncertainty in relative pose measurements. Moreover it can be naturally extended to different communication technologies (bluetooth, RFId, wifi, etc.) regardless their sensing range. The proposed approach is validated through experimental test.

Autonomous exploration under uncertain robot position requires the robot to plan a suitable motion policy in order to visit unknown areas while minimizing the uncertainty on its pose. The corresponding problem, namely active SLAM (Simultaneous Localization and Mapping) and exploration has received a large attention from the robotic community for its relevance in mobile robotics applications. In this work we tackle the problem of active SLAM and exploration with Rao-Blackwellized Particle Filters. We propose an application of Kullback-Leibler divergence for the purpose of evaluating the particle-based SLAM posterior approximation. This metric is then applied in the definition of the expected gain from a policy, which allows the robot to autonomously decide between exploration and place revisiting actions (i.e., loop closing). The technique is shown to enhance robot awareness in detecting loop closing occasions, which are often missed when using other state-of-the-art approaches. Results of extensive tests are reported to support our claims.

Different solutions to the robot control problem by using only position measurements have been studied in literature. The performances of some linear and nonlinear observers previously proposed are discussed in this paper, on the basis of the experimental results obtained in the control of a SCARA two-link manipulator with two revolute joints. The comparison is carried out by using the observed velocities to implement a linear state feedback algorithm for the robot control. The obtained results show that different factors (e.g. the required task, the available knowledge of the robot dynamic model) may influence the choice of a particular observer.

Common robot calibration procedures use least-squares (LS) techniques to obtain estimates of the identifiable parameters. The "quality" of the resulting estimates depends significantly upon the used excitation input. The search for the best excitation trajectory is usually posed as a nonlinear path optimization problem aimed at optimizing suitable measures of the LS normal equations matrix. In this paper a parametrization of the class of reference joints trajectories is introduced and a solution framework based on genetic evolution is proposed. The efficiency of the method is illustrated by experimental tests on a SCARA two-link manipulator. Issues related to data acquisition and signals reconstruction and filtering are also discussed.

The authors present an algorithm for hybrid force-position control of a robot manipulator using the operational space formulation approach. The proposed scheme modifies and integrates a previous algorithm due to O. Khatib (1987) and to Khatib and J. Burdick (1986), to achieve an exact decoupling of the force and position control along the task subspaces. The theoretical results presented demonstrate the viability of this approach, which was confirmed by simulation results. These provided evidence that the decoupling between force and position was exact, with no noticeable increase in the required computational burden.>