ABSTRACT. This work introduces a mobile robotic platform developed for autonomous navigation in wooded environments and for automatic estimation of the location of individual trees and their trunk diameters. The proposed system combines LiDAR measurements and camera images within a framework that leverages SLAM together with deep learning for trunks recognition. The proposed navigation and mapping approach is tested in a wooded area near Udine (Italy), using a skidsteered mobile robot. Experimental results demonstrate that the robot is capable of navigating effectively avoiding obstacles while generating forest maps enriched with tree trait information.

ABSTRACT. This paper presents simulations of Search andRescue (SaR) using multiple Unmanned Aerial Vehicles (UAVs)equipped with ARVA (Appareil de Recherche de Victimesd'Avalanche) beacons that use Extremum Seeking Control(ESC) to reduce the time operation of rescue avalanche victims.ESC, a real-time optimization technique, enables UAVs toautonomously navigate toward the signal source withoutrequiring a detailed system model. The simulations comparerescue times using single and multiple UAVs configurations in a100m x 200m area. Results show that using four UAVs reducesthe victim detection time from 230 seconds (single UAV) to 160seconds, demonstrating improved efficiency. This work includesmathematical modeling of the magnetic field generated byARVA transmitters and validates the ESC strategy throughsimulations.

ABSTRACT. Social robots must perceive and navigate safely around humans in confined spaces. This work employs DINOv2 ViT-B/14 to extract visual features from RGB images captured by a quadruped robot in indoor environments by considering two input strategies (full-image and sliding-window), and by evaluating the model in configurations with and without token registers to assess their influence on feature extraction. Results demonstrate that the inclusion of registers, regardless of the input approach, improves spatial coherence, suggesting DINOv2 as a reliable backbone for future perception and navigation modules.

ABSTRACT. To meet rising quality standards in plant production while addressing increasing environmental and societal concerns, the agricultural industry faces growing pressure to develop innovative solutions. Thus, novel approaches are essential to address sustainable practices, ongoing labor shortages and improve tasks throughout the life cycle of agricultural products. In this context, our project focuses on the design, control, and preliminary validation of a robotic system for precision agricultural processes, guided by stakeholder and end-user feedback. The platform integrates three modules: navigation, which employs path planning and obstacle avoidance; perception, to detect grapes and identify their positions within the vineyard; and manipulation, to perform tasks such as inspection and precise treatment application.

ABSTRACT. The shortage of skilled operators and increasing safety requirements in harsh environments like construction sites, forests, and mines demand innovation in construction machinery. Simplifying control and enhancing machine intelligence are essential to address these challenges. Inspired by robot-avatar concept, which enables remote human activity via intuitive interfaces, we explored how to improve construction machine usability, making them safer and accessible to untrained users. Motivated by this, we developed an industrial Gigabot architecture composed by cartesian control and teleoperation technologies. The system was validated operating a crane remotely at the Bauma 2025, the world’s leading construction machinery fair. Thanks to the positive feedback received by crane operators and interest for potential application, the industrial partner is currently industrializing it.

ABSTRACT. This work introduces an autonomous robotic measurement platform, based on a quadruped robot, designed to perform systematic 5G spectrum measurements and area coverage analysis. Fifth-generation (5G) mobile networks have increased the complexity of radio propagation: unlike previous generations, 5G signals show strong spatial and temporal variability, making them difficult to characterize comprehensively with static, manual surveys, thus requiring new measurement approaches.

ABSTRACT. This paper presents a novel cable-driven hyperredundant robot featuring flexure hinges-based compliant modules. The 12-degrees-of-freedom fully 3D-printed manipulator employs six universal joints with Bowden cable actuation for decoupled control. Two backbone-based inverse kinematics algorithms are developed and compared against traditional methods. Experimental validation using motion capture demonstrates joint accuracy below 8.59 deg and the capability of the system to carry 200 g payload. The heuristic algorithm achieves 0.41 mm position accuracy with less than 2 ms cycle time, significantly outperforming Jacobian-based approaches.

ABSTRACT. Visual active tracking has become an increasingly prominent topic in robotics because of its central role in applications such as assistive robotics and surveillance. Unlike passive tracking, active methods tightly couple perception and control: the system not only detects the target but also actively maneuvers the platform to keep it in view. While much of the literature focuses on ground robots, solutions for aerial platforms remain scarce. To address these issues, in this extended abstract we present our ongoing work on visual active tracking for micro aerial vehicles, focusing on the comparative evaluation of two complementary design strategies. The first adopts a modular approach, in which perception and control components are developed separately and then integrated into a pipeline; the second follows an End-to-End approach, where a single deep policy directly maps visual observations to control commands through reinforcement learning. By contrasting these paradigms, our goal is to assess their relative strengths, limitations, and applicability to real-world deployment.

ABSTRACT. In viticultural environments, trajectory planning for a robotic arm is particularly challenging due to the complex structure of grapevines. We propose a method for precise engagement of a robotic manipulator’s shears with grapevine canes during automated winter pruning. The approach utilizes artificial potential fields to guide the robot toward the target, while employing vortex fields to safely navigate the shears around obstacles detected from a 3D point cloud. Experimental results demonstrate that the method successfully guides the robot to the desired cane while avoiding collisions between the blades and the plant, enabling accurate and safe pruning operations in realistic vineyard conditions.

ABSTRACT. We present an online semantic object mapping system for a quadruped robot operating in real indoor environments, turning sensor detections into named objects in a global map. During a run, the mapper integrates range geometry with camera detections, merges co-located detections within a frame, and associates repeated detections into persistent object instances across frames. Objects remain in the map when they are out of view, and repeated sightings update the same instance rather than creating duplicates. The output is a compact object layer that can be queried (class, pose, and confidence), is integrated with the occupancy map and readable by a planner. In on-robot tests, the layer remained stable across viewpoint changes.

ABSTRACT. Autonomous navigation in unstructured outdoor environments presents significant challenges. While deep Reinforcement Learning (RL) has shown promise, it often faces difficulties in generalization and data efficiency. To address these issues, we evaluate the performance of four deep RL algorithms (PPO, A2C, SAC, and TD3) for a point-goal navigation task within MIDGARD, our photorealistic simulator built on Unreal Engine. By comparing success rates and reward progression, we demonstrate that PPO significantly outperforms the other algorithms. Our results indicate that on-policy methods are better suited for this navigation task, whereas off-policy methods struggle due to inefficient exploration and policy updates

ABSTRACT. Hand exoskeletons can restore functions in individuals with hand motor impairments, supporting everyday life. This work presents a preliminary development of a modular, rigid finger exoskeleton, custom-designed for a specific user by using 3D scanning and additive manufacturing technologies. A linkage mechanism was synthesized and 3D-printed, integrating custom cuffs and a linear actuator. Performance analysis demonstrated that the exoskeleton module produces a natural motion, closely matching desired finger kinematic data, and highlights the potential of personalized solutions for effective hand assistance.

ABSTRACT. Research on hand gesture recognition has explored inertial, optical, and EMG sensors, but wearable and efficient solutions remain limited. This work presents a glove with soft differential capacitive sensors and a microcontroller for real-time classification of four hand grasps, achieving 87.5% accuracy with a dense neural network on the Arduino Nano 33 BLE Sense.

ABSTRACT. Gait analysis is a fundamental component of clinical and rehabilitation practice; however, traditional marker-based systems are costly, technically demanding, and restricted to con- trolled laboratory environments. While markerless approaches offer greater accessibility, they often fail to provide biomechanical outputs without the integration of complex modeling frameworks, limiting their applicability in clinical settings. This study presents a neural network pipeline that derives clinically meaningful joint kinematics directly from parametric body model data, thereby eliminating the need for marker-based acquisition or biomechan- ical modeling software. The proposed method is intended to be lightweight, adaptable, and compatible with data from commonly available sensing devices, enabling its use beyond laboratory conditions. This approach could support more accessible and efficient gait assessment in everyday clinical practice.

ABSTRACT. Robot-aided rehabilitation enables tailored, repeatable training for patients with motor impairments, but current adaptation strategies often rely on pre-set rules and may not reflect therapists’ real-time decisions. This study analyzed motor, physiological, and subjective patient data from eight orthopedic patients to identify the factors guiding therapists’ adjustments of assistance levels. Results show that therapists primarily rely on subjective feedback of exertion and pain, while kinematic and kinetic metrics have limited influence. These findings support integrating patient-reported measures into robotic platforms to enhance adaptive assistance and align therapy with clinical judgment.

ABSTRACT. Neurocognitive impairments encompass a spectrum of conditions that lead to deterioration of cognitive functions, with symptoms worsening over time, and impose significant physical, psychological, social, and economic burdens on patients, caregivers, and society. Early cognitive training and continuous monitoring have proven beneficial in slowing decline and supporting daily functioning. This project originates from the need for innovative technological tools to train, monitor, and assess the cognitive abilities of individuals with dementia It introduces a tablet-based advanced digital platform, designed according to a user-centered design approach, for cognitive training and assessment in individuals experiencing cognitive decline, that delivers gamified exercises across the six cognitive domains defined in the Diagnostic and Statistical Manual of Mental Disorders guidelines. The system integrate adaptive difficulty, performance monitoring, and the potential for multiplayer interaction, to promote engagement and provide longitudinal data for clinicians and caregivers. A preliminary beta version is under evaluation to assess usability, engagement, and clinical relevance.

ABSTRACT. The adoption of robotic technologies and multimodal monitoring systems has emerged as a promising approach in rehabilitation, offering precise assistance, objective measurement, and the possibility to design adaptive and personalized protocols addressing both motor and cognitive functions. In this work, a cognitive rehabilitation platform is presented, based on TIAGo service robot, able to operate in joystick mode for compliant human–robot interaction or in monitoring mode for user state tracking. The system combines robotic kinematics, physiological signals and affective features, orchestrated by a finite-state machine to ensure safe interaction. Preliminary tests with healthy volunteers demonstrated the feasibility of integrating robotic interaction with multimodal assessment. Future developments will extend validation to patient populations and implement real-time adaptation of rehabilitation tasks based on arousal and valence estimates, paving the way for personalized and adaptive clinical protocols.

ABSTRACT. This study presents the preliminary development of AGLAIA, a unified platform for global assessment integrating sensorimotor and cognitive domains. The robotic device EDUSA® PRO-R, GlovETA with EMG sensorized sleeve and the cognitive platform Nu!Reha are integrated into a single Human-Machine Interface. A protocol was developed and tested with 11 unimpaired participants to validate the platform, establishing the foundations for a subsequent phase focused on data driven personalized training.

ABSTRACT. Tactile and orientation information are very important to improve robotic manipulation capabilities in several contexts such as healthcare, industrial, logistic and domestic. This abstract aims at presenting a multi-modal sensing system that optimize accuracy through the integration of tactile sensors, an Inertial measurement unit and a connector for a Time-of-Flight pre-touch sensor. While tactile data enables closed-loop control during grasping operations, the measurements from the inertial unit are elaborated by state-of-the-art sensor fusion algorithms to obtain the sensor orientation in the 3D world.

ABSTRACT. Wearable haptic technologies are emerging as promising tools for home-based therapy. We present a robotic patch capable of delivering controlled mechanical stimulation to human skin through compact actuation, compliant design, and integrated sensing. Experimental tests confirmed its ability to reproduce clinically relevant indentation and pressure levels, while morphological analysis demonstrated broad applicability across body regions. These findings suggest that wearable mechanotherapy can provide a safe and adaptable approach to support rehabilitation.

ABSTRACT. This study presents a learning-based framework for multi-touch characterization of the SoftMag sensor. A combined classification and regression approach is employed to simultaneously identify contact regions and estimate corresponding interaction forces. The proposed method enhances tactile resolution and supports robust perception for integration into compliant robotic systems.

ABSTRACT. The sense of touch is essential for dexterous and safe robotic object manipulation. The same holds in teleoperation, where a major challenge is still how to deliver intuitive tactile feedback from the robot to the human operator. This work presents a wireless, real-time system that integrates an array of Fiber Bragg Grating tactile sensors in a custom robotic gripper, with a wearable haptic interface. Multiple sensor signals were combined and transformed into intuitive force feedback, allowing a remote user to perceive manipulation interactions of the robot with an object in a quasi-natural manner.

ABSTRACT. This study introduces a soft inflatable actuator designed to provide rich, multimodal sensory feedback, including both pressure and vibration. The actuator, weighing less than 2 g and with a thickness of 0.4 mm, is a significant improvement over previous designs. Mechanical characterization demonstrates the actuator's capability to produce high-bandwidth vibrations up to 200 Hz and high forces of 28 Newtons at 60 kPa. Psychometric tests conducted with 14 able-bodied individuals and three transradial amputees show performance comparable to state-of-the-art invasive and non-invasive solutions. Furthermore, the actuator successfully conveys different artificial roughness sensations to able-bodied individuals. In a classification task, amputees achieved an overall accuracy of 73.3%, with touch and pressure being the dominant among the elicited sensations. These results highlight the effectiveness, versatility, and lightweight nature of the proposed soft feedback actuator, showing its potential for integration into robotic systems for feedback restoration and augmentation.

ABSTRACT. Tactile sensing provides information about touch-based mechanical interactions with the environment. The intent of this work is to review our recent research outputs and outline future perspectives. In particular, we present two complementary approaches for processing tactile data acquired through an artificial skin based on piezoelectric polymers.

ABSTRACT. In this work, an algorithm for estimating the position and Gaussian-like shape of an unknown number of defects on a given flat surface is developed. A natural behavior for exploring/exploiting possible defects using the ergodic control theory for trajectory planning has been achieved. Informative tactile measurements along with the end-effector/sensor position related have been used to provide a posterior belief using the Gaussian Mixture Model technique, and a Kullback-Leibler divergence-dependent constraint has been designed to enhance the accuracy of the estimation.

ABSTRACT. This paper presents an extension of Thunder Dynamics for elastic joint systems. Thunder Dynamics provides automated tools for generating dynamic models and parameter identification regressors through both high-level and low-level interfaces. The library enables users to create standalone C++ libraries with Python bindings and supports symbolic parameter manipulation for adaptive control implementation. We demonstrate potential applications to variable stiffness actuators characterized by nonlinear elastic couplings modeled as polynomials with odd exponents. The framework addresses the critical need for standardized tools in robotics research and development, offering both computational efficiency and extensibility for handling complex robotic systems, including elastic joint mechanisms.

ABSTRACT. In this work, we present a direct-drive approach applied to an haptic thimble based on soft actuated belts. Building on a previous gearmotor-based design, the method leverages direct transmission to enhance linearity and to extend the output bandwidth of the device, thereby enriching the rendering of dynamic tactile feedback. While the approach reduces the maximum achievable force, it enables wide-bandwidth interaction in a thin and lightweight form factor. We validate the device in a virtual reality manipulation experiment, where physical simulation requires careful pick-and-place of virtual cubes. We also propose a proof-of-concept implementation in a teleoperation setup, where the thimble is coupled with a robotic hand equipped with sensitive pressure-based fingertip sensors. Results in VR highlight the effectiveness of the feedback in reducing the average virtual indentation and improving repeatability of the fine manipulation task.

ABSTRACT. Few linkage-driven devices achieve finger Shape Adaptivity (SA) because of the difficulties in providing enough degrees of freedom during grasping without increasing the overall encumbrance. This work proposes a method to provide four-bar linkage fingers with SA by substituting one rigid link with a compliant element. Compliance was obtained by inserting into a Nylon rigid link a compliant domain in TPU 95A using Additive Manufacturing Multi-material printing. Different interlocking interfaces and printing parameters were tested to optimise link compliance. Compression tests identified the best parameter combination, and cyclic tests confirmed the stable behaviour of TPU. These results enable the development of linkage-driven prosthetic fingers with SA, overcoming current limitations.

ABSTRACT. We present a transhumeral prosthesis with Variable Stiffness Actuators (VSAs) to mimic the adjustable compliance of human joints. The modular design accommodates different users' residual limb shapes and biological control signals. Our design draws on biological compliance for natural interaction with the environment, aiming to improve prosthesis embodiment and users' social interactions. Moreover, embedded elastic elements enhance safety, enable more natural movement, and enhance versatility in daily tasks.

ABSTRACT. Lack of tactile feedback remains a major limitation in prosthetics, restricting manipulation accuracy and increasing cognitive effort during daily activities. For this reason, modern artificial hands must be endowed with force or tactile sensors. A key question is whether additive manufacturing plastic structure can themselves host load sensing elements. Here we investigate the integration of semiconductor strain gauges into an additively manufactured proximal phalanx of a prosthetic finger as a structural sensing approach for force estimation. We propose a methodological framework to identify the optimal sensor configuration, ensuring independence from the contact point as well as robustness to finger posture variations. The phalanx was further refined with cavities that concentrate strain in the sensing regions, enhancing measurement sensitivity and overall system effectiveness. Finite element simulations guided the structural optimization and assessed sensor performance under representative loads. Simulation results demonstrate that the introduction of cavities can increase strain sensitivity by a factor of up to 2.5, while reducing crosstalk between sensing axes and improving robustness across different grasp types. This confirms that structural modifications enabled by additive manufacturing can significantly enhance the effectiveness of embedded strain-based sensing. Ongoing efforts focus on the experimental validation, with the potential to establish structural additive manufacturing sensing as a potential solution for affordable, personalized prosthetic hands with integrated tactile feedback.

ABSTRACT. This work presents the preliminary validation of an adaptive control strategy for a hand-wrist prosthesis, designed to coordinate wrist motion with grasping. The proposed adaptive approach modulates wrist rotation speed in real time based on tangential forces and slip detection to improve motion accuracy, and was compared against a constant velocity control strategy. Experimental validation with eight participants showed a significant reduction in errors when using the adaptive modality, demonstrating that adjusting wrist velocity based on hand–object interaction improves performance in dynamic manipulation tasks.

ABSTRACT. Transcutaneous Electrical Nerve Stimulation (TENS) represent a promising approach to provide sensory feedback in lower limb prostheses, since it allows the elicitation of homologous and somatotopic sensation in a non invasive way through superficial electrodes that stimulate the underlying nerves. Since ground unevenness is one of the most influential environmental factors affecting the risk of falls in amputees, restoring the ability to discriminate floor conditions could play a crucial role in improving prosthesis usability and safety. However, the potential of TENS to convey different sensory cues according to terrain characteristics has not yet been explored. This study aimed to develop and test a novel encoding algorithm for lower limb sensory feedback restoration enabling the recognition of different terrain textures. The system performance was evaluated on five healthy participants for discrimination of three terrain textures. The proposed approach led to a correct classification of grass, stones and tiles in 88%, 81% and 79% of the cases with an overall accuracy of 83% outperforming literature results.