Programma tecnico dettagliato poster session I-RIM 2022

7 Ottobre, ore 10:00 – 13:00

Istituto Superiore Antincendi – Sala Caravaggio


FR P1 1

An Approach to Exploit Compensatory Motions in Upper-Limb Prostheses Control

  Maddalena Feder, Giorgio Grioli, Manuel Giuseppe Catalano and Antonio Bicchi
  ABSTRACT. Among the recent investigations in upper-limb prostheses, the research still focuses on exploring new control solutions to reduce the user’s mental fatigue and improve the control’s robustness and intuitiveness. Some studies present solutions to close the control loop by using compensatory motions as error indexes. A previous relation is established to associate a predefined compensation motion to a given prosthetic degree of freedom. We developed an approach to avoid this step and directly correlate the human motions to as many prosthetic joints as needed. The implemented algorithm is tested in Simulink for a simulated trans-humeral amputee.

FR P1 2

Bioinspired microhooked patches for targeted delivery of molecules in plant leaves

  Isabella Fiorello and Barbara Mazzolai
  ABSTRACT. In the present work, we mimic the strong shear-dependent leaf attachment of the hook-climber Galium aparine to propose new patches with micropatterned hooks for targeted delivery of molecules in leaf tissues. We first built biodegradable and dissolvable isomalt-made micro-hooked patches using high-resolution micromanufacturing techniques including two-photon lithography. Secondly, we characterize the mechanical properties of plant leaves and the attachment forces of the micro-hooked patches to leaf surfaces. Lastly, we tested the micro-hooked patches for in situ release of fluorescein molecules in plant vascular tissues, producing biodegradable environmental-friendly machines. This research highlights the potential to use plant-like sustainable miniaturized machines to cure plants, reducing the use of pesticides and preserving natural ecosystems.

FR P1 3

Comparison of Haptic Strategies for Human Guidance

  Tommaso Lisini Baldi, Gianluca Paolocci and Domenico Prattichizzo
  ABSTRACT. This paper compares two human guidance approaches mediated by haptic stimulation. Motion Guidance consists in delivering step-by-step instructions to human users, that respond to commands with specific actions. The Sensory Augmentation approach enriches the human with the knowledge necessary to complete the task using a self-selected strategy. The two approaches were evaluated in a collaborative scenario with couples of participants carrying a bulky object under the sole guidance of haptics. Stimuli were delivered by a haptic belt according to three haptic policies. According to results, both approaches performed well when participants were aware of the direction to the target during the whole trial, supporting the idea of using haptics to convey environmental information to the users.

FR P1 4

Human-Robot Collaboration Using Fuzzy Adaptive Virtual Fixture Method for Dental Implant Surgery

  Mohammad Hossein Hamedani, Fan Shao, Emanuele Vaia, Alessandro Cuozzo, Luca Ramaglia, Bruno Siciliano and Fanny Ficuciello
  ABSTRACT. The purpose of this work is to develop a methodology to improve human-robot collaboration for robot-aided dental implant placement. In this study, a human-robotic implant system (HRIS) is designed according to a hand-guiding control to increase the accuracy and stability of osteotomy drilling based on the surgeon’s decision, and robot motion during the implant placement. The proposed method is able to guide the surgeon’s hand according to the pose of the desired placement. To guide and modify the pose of the surgeon’s hand, the virtual fixture method is used as the main control approach. To verify the performance of the introduced method, the KUKA MED robot is used to perform the dental implant placement using the presented approach on a phantom head with a 3D jaw bone model. Additionally, the results between free-hand drilling and HRIS controlled drilling according to the apical center and head center of the implant placement are compared to evaluate the performance of the introduced method.

 FR P1 5

Human recognition for resource-constrained mobile robot applied to Covid-19 Disinfection

  Andrea Mattia Garavagno, Daniele Leonardis and Antonio Frisoli
  ABSTRACT. The global COVID-19 pandemic has stimulated the use of disinfection robots: in September 2021, following a European Commission’s action, 200 disinfection robots were delivered to European Hospitals. UV-C light is a common disinfection method, however, direct exposure to UV-C radiation is harmful and disinfection can be operated only in areas strictly forbidden to human personnel. We believe more advanced safety mechanisms are needed to increase the operational flexibility and safety level. We propose a safety mechanism based on vision and artificial intelligence, optimised for execution on mobile robot platforms. It analyses in real-time four video streaming and disables UV-C lamps when needed. Concerning other detection methods, it has a relatively wider and deeper range, and the capability to operate in a dynamic environment. We present the development of the method with a performance comparison of different implementation solutions, and an on-field evaluation through integration on a mobile disinfection robot.

FR P1 6

A Smart Robotic Platform for Monitoring Food Assumption in Elderly Patients

  Francesco Scotto di Luzio, Nevio Luigi Tagliamonte and Loredana Zollo
  ABSTRACT. Malnutrition is a prevalent comorbidity in hospitalized elderly patients, often caused by dysphagia and potentially increasing the risk of prolonged hospitalization and mortality. Such patients must follow a specific diet and often carry out speech exercises with the constant presence of a clinical supervisor. However, there are several difficulties in monitoring patients and also obtaining complete information on the intake of food. Several robotic solutions have been presented in the literature to support the patient and motivate him/her during food assumption, but no technological solutions have been proposed for elderly patients at risk of undernutrition. This paper proposes an approach for using a service robot for the monitoring of the required energy intake to avoid weight and muscle mass loss and identifing any event of cough or oxygen desaturation in hospitalized geriatric patients.

FR P1 7

Experimental Psychophysical Characterization and Validation of Sound Exciter Vibrotactile Actuators in Haptics

  Alessia Silvia Ivani, Manuel G. Catalano, Simone Fani, Giorgio Grioli and Antonio Bicchi
  ABSTRACT. Vibrotactile actuators, such as voice coils, are effective, non-invasive solutions commonly used in haptics to render tactile information. Among voice coils, sound exciters offer several advantages to wearable haptics through superior economic affordability, compact size, and commercial availability. This study explores the feasibility of using disk-shaped sound exciters actuators for wearable haptic applications, bench-marking them against traditional voice coil cylindrical actuators. Seven participants took part in contact and roughness discrimination experiments. The qualitative and quantitative study investigates the Just Noticeable Difference (JND) and Absolute Threshold (AT), exploiting standard psychophysical methods. The proposed protocol effectively compared the performance of the two actuators. Our preliminary results show that sound exciters may be a viable alternative to more commonly used vibrotactile actuators for practical haptic applications.

FR P1 8

Optimal electromyographic sensing for whole-body muscular activity estimation

  Marco Baracca, Giuseppe Averta and Matteo Bianchi
  ABSTRACT. Recording the level of muscles activation while human perform daily-living activities is a key factor to obtain a comprehensive evaluation of the biomechanical state. This type of analysis is especially important in work environments to identify and then avoid possible risky behaviours which could lead to work-related musculoskeletal disorders. Usually, muscular activation is recorded via wearable electromyographic EMG sensors. However, to achieve a whole-body state estimation a large number of sensing elements is necessary. This leads to uncomfortable and very expensive setups that prevent their adoption for monitoring daily working activities. To overcome this problem, we propose a solution to provide a reliable estimation of muscular activation from a limited number of EMG recordings. Our method exploits the covariation patterns between muscular activation signals to complement the recordings coming from a reduced set of optimally placed sensors, minimizing the estimation uncertainty. We tested this approach with a dataset containing EMG data from 10 different subjects. We were able to reconstruct the temporal evolution of 10 whole-body muscular activations with only 7 sensor elements achieving a maximum normalized estimation error of 13%.

FR P1 9

Insights on Oscillations and Impulses in Variable Impedance Control

  Giorgio Simonini, Rachele Nebbia Colomba, Chiara Sammarco, Mathew Jose Pollayil, Franco Angelini and Antonio Bicchi
  ABSTRACT. Operational-space impedance control allows to assign a mass-damper-spring (MDS) behavior to the end effector of a robotic manipulator. This aids resilience when interactions with the environment or humans are expected. While changing the virtual mass is not always practicable since it requires using precise force/torque sensors, varying the stiffness and damping can be useful to intuitively modify the response of the robot to external perturbations. In this paper, considerations are made on MDS systems regarding oscillations and impulse responses of the system when its impedance is varied. These observations are then validated in simulation using a two degrees of freedom (DoFs) robotic arm.

FR P1 10

Behavioural Maps for Assisted Navigation

  Placido Falqueto, Alessandro Antonucci, Luigi Palopoli and Daniele Fontanelli
  ABSTRACT. In this paper, we propose a new paradigm for robot-assisted navigation: human behaviour is accepted if, at any time, it follows what other humans would do in the same situation. This is achieved through the combination of machine learning and adaptive control. The performed experiments validate the feasibility of this idea.

FR P1 11

Human-Robot Role Arbitration with Cooperative Game Theory

  Paolo Franceschi
  ABSTRACT. This work studies the role-arbitration between a robot and a human during physical Human-Robot Interaction. The system is modeled as a Cartesian impedance, with the human and the robot interacting by applying two separate external forces. A reformulation of the problem as a Cooperative Differential Game allows addressing the bargaining problem by proposing a solution depending on the human interaction force, interpreted as the will to lead or follow. This defines the arbitration law and assigns the role of leader or follower to the robot. Experiments show the feasibility and capabilities of the proposed control in managing the human-robot arbitration during a shared-trajectory following task.

FR P1 12

A data-driven approach to human-robot co-manipulation of soft materials

  Giorgio Nicola, Enico Villagrossi and Nicola Pedrocchi
  ABSTRACT. Human-robot co-manipulation of large but lightweight elements made by soft materials is a challenging operation that presents several relevant industrial applications. This paper proposes using a 3D camera to track the deformation of soft materials for human-robot co-manipulation. Thanks to a Convolutional Neural Network (CNN), the acquired depth image is processed to estimate the element deformation. The output of the CNN is the feedback for the robot controller to track a given set-point of deformation.

FR P1 13

Towards the Use of Programming without Coding for Navigation Tasks

  Valeria Sarno, Gianluca Lentini, Antonio Bicchi and Lucia Pallottino
  ABSTRACT. Learning from Demonstration is a powerful method that allows robots to acquire skills from humans through imitation. There are several strategies in the literature for teaching robots manipulation tasks. However, the aim of this paper is to propose a preliminary study on the possibility of teaching a mobile robot navigation tasks using a single demonstration. To this end, we employed a framework, based on Dynamic Movement Primitives, which has already been proven effective in manipulation tasks. We tried to adapt it to our scenario, taking into account the differences between the two types of tasks. Our approach was validated in a simulated environment with a mobile base equipped with a 2-D laser scan. Preliminary results seem to confirm that it is possible to extend that framework to navigation tasks.

FR P1 14

Robotic Non-prehensile Object Transportation

  Mario Selvaggio, Viviana Morlando and Fabio Ruggiero
  ABSTRACT. This document revises the latest results related to robotic non-prehensile object transportation. The problem consists of a robotic arm transporting an object along the desired trajectory on a tray, guaranteeing a sticking behaviour and other constraints. The solution in [1], where an optimal control problem has been devised, is revised together with the case in [2], where a quadruped/mobile robot is considered. A model predictive control approach is also described to solve the same problem.

FR P1 15

Learning Trajectory Tracking for Underactuated Compliant Arms

  Michele Pierallini, Franco Angelini and Manolo Garabini
  ABSTRACT. Trajectory tracking is a classic control theory topic that has received in-depth research in the literature. However, dealing with compliant arms that is underactuated makes the issue more difficult. Compliant systems frequently exhibit difficult-to-model dynamics in addition to their underactuation. To prevent a severe modification of the robot elasticity, the feedback components should be limited. In this letter, we use an iterative learning controller to solve the trajectory tracking problem. The presented control law mixes feedforward and feedback terms. The feedforward component tracks the desired trajectory raising the robot to one equilibrium, and the feedback term stabilizes the equilibrium. We investigate the closed-loop stiffness variation. Finally, we simulate an underactuated compliant arm to verify the suggested technique.

FR P1 16

Visual Active Exploration and Tracking: a Deep Reinforcement Learning Approach

  Alberto Dionigi, Alessandro Devo, Leonardo Guiducci, Paolo Valigi and Gabriele Costante
  ABSTRACT. Visual tracking approaches have recently gained significant attention from the research community. The majority of the state-of-the-art methods propose passive visual tracking solutions, i.e., the camera cannot take actions to change its field of view in order to track the target. The objective of active approaches, instead, is to develop trackers capable of computing motion controls to maintain visual contact with the target. However, these strategies are still scarcely explored and most of those assume that the target to be within the tracker field of view at the beginning of the experiment. To remove this limitation, in this work we propose a novel deep reinforcement learning-based approach that is capable of exploring the surrounding environment, find the target and track it. We design and run different experiments to show the effectiveness of our approach and compare it with respect to current state-of-the-art (SotA) methods.

FR P1 17

FASTDLO: Towards Real-Time Perception of Deformable Linear Objects

  Kevin Galassi, Alessio Caporali and Gianluca Palli
  ABSTRACT. In this paper is presented an approach for fast and accurate segmentation of Deformable Linear Objects (DLOs) named FASTDLO. The perception is obtained from the combination of a deep convolutional neural network for the background segmentation and a pipeline for the dlo identification. The pipeline is based on skeletonization algorithm to highlights the structure of the DLO and a similarity-based network to solve the intersection. FASTDLO is trained only on synthetically generated data, leaving real-data only for evaluation purpose. FASTDLO is experimentally compared against DLO-specific approach achieving better overall performances and a processing rate higher than 20 FPS.

FR P1 18

Soft Collaborative Gripper for Dressing Assistance

  Mihai Dragusanu, Sara Marullo, Monica Malvezzi, Gabriele Maria Achilli, Maria Cristina Valigi, Domenico Prattichizzo and Gionata Salvietti
  ABSTRACT. We present the DressGripper, a collaborative gripper designed for safe interactions in wearing operations. The DressGripper accounts for the two main goals that have to be achieved by a robotic system designed for helping people to get dressed: i) to be intrinsically safe and ii) to keep firmly the cloth during the dressing operations. The DressGripper addresses these issues by combining a compliant and safe structure with an additional magnetic actuation at the fingertips. This solution allows a soft interaction with the robot and guarantees the grasping tightness. Preliminary experiments show the DressGripper suitability in robotic dressing assistance scenarios.

FR P1 19

Human-Robot Collaboration during the Execution of Structured Co-manipulation Tasks

  Jonathan Cacace, Riccardo Caccavale, Alberto Finzi, Riccardo Grieco and Vincenzo Lippiello
  ABSTRACT. In this work, we consider human-robot collaboration during the execution of structured co-manipulation tasks to be interactively accomplished exploiting the human physical guidance. In this scenario, human interventions are continuously assessed by the robotic system to infer whether the human guidance is consistent with respect to the shared plan. On the other hand, the estimated intentions associated with the human interventions are also exploited by the robot to on-line regulate its cooperative behavior in the context of the shared task. For this purpose, the collaborative robot can adjust tasks/subtasks, or motions, while regulating its compliance with respect to the operator’s guidance

FR P1 20

Automated Design of Add-Ons for Soft Hands

  Valerio Bo, Enrico Turco, Maria Pozzi, Monica Malvezzi and Domenico Prattichizzo
  ABSTRACT. Soft robotic hands allow the exploitation of hand-object-environment interactions, but their capabilities can still be limited in cluttered or narrow spaces. In this abstract, we propose enhancing the versatility of soft grippers by adding passive components to their structure without completely altering their design and control. These add-ons are prototyped through a data-driven automated design procedure and their effectiveness has been proved with experiments using two different hands.

FR P1 21

Robotic Waste Sorting for urban recycling

  Emanuele Menegatti, Alberto Bacchin, Nicola Castaman, Stefano Tonello and Nicola Carlon
  ABSTRACT. Our litter contains a large number of raw materials which are wasted. The different materials must be separated to be economically enhanced and today technology does not allow you to do this efficiently. Most often it is necessary to use personnel who work in unhealthy conditions to recover only a portion of the recoverable material. Our project aims to create a robotic sorting system for sorting the different materials in our waste. This is achieved thanks to advanced computer vision and robot manipulation techniques. Waste sorting is not only a challenging technical problem, but it is a problem with several implications. Environmental implications, because of waste disposal. Social implications, due to the harsh working conditions of workers and the high turnover of staff. Economic implications, to make profitable recycling and recovery of waste. In this paper, we present the basic building blocks on which we are creating our solution to the robotic waste sorting problem.

FR P1 22

An Anthropomorphic Aerial Manipulator equipped with a Testbed for Indoor Experiments

  Fanyi Kong, Simone Monteleone, Giorgio Grioli, Manuel Giuseppe Catalano and Antonio Bicchi
  ABSTRACT. During interventions in hazardous environments, debris often restricts the accessibility of ground robots. When this occurs, Unmanned Aerial Manipulators (UAMs) can intervene, avoiding all of these obstacles. However, when UAMs enter and interact with such environments, designers must ensure that they possess high stability and robustness. This work presents a teleoperated aerial manipulator for intervention in unstructured environments and a testbed to perform safely indoor experiments for evaluating the stability performance and manipulation tasks. The robot comprises an aerial platform, two arms that terminate with two Pisa/IIT Softhands, and a head equipped with stereo cameras. The testbed constrains the drone to move in a limited space and avoids that, in case of control loss, the drone falls or rambles in an indoor space, which may cause damage to the robot or humans.

FR P1 23

A Cloud Architecture for Emotion Recognition in Human-Robot Interaction Based on the Appraisal Theory

  Marco Demutti, Vincenzo D’Amato, Carmine Recchiuto, Luca Oneto and Antonio Sgorbissa
ABSTRACT. This work proposes a cloud system, structured as a set of REST API endpoints, for online human emotion recognition in spontaneous human-robot verbal interaction. Based on the appraisal theory of emotion, the system acquires data about the person’s expected appraisal of a given situation, depending on their needs and goals, and combines it with sensory data, such as facial expressions, angles of the head, and gaze of the person, and distance between the person and the robot. The whole set of data is used to infer the emotional state of the person during the interaction through a Random Forest classifier, trained for binary classification (i.e., positive vs. negative emotions). Results confirmed that using both sources of data led to a performance improvement both in the K-fold and in the Leave One Person Out scenarios.

FR P1 24

TERglove: a modular low-cost IMU-based Data Glove

  Valerio Belcamino, Alessandro Carfì, Mohamad Alameh and Fulvio Mastrogiovanni
ABSTRACT. For humans, the hands represent the most important medium of interaction with the environment. In fact, their sophisticated kinematic structure and perception abilities allow to hold objects, perform gestures and perceive contact forces. The development of tools capable of detecting this information has become increasingly important in recent years given the implications in various fields ranging from medicine to robotics.




8 Ottobre, ore 9:00 – 12:00

Istituto Superiore Antincendi – Sala Caravaggio

SA P2 1

Experimental microbiological analysis of a UV-based sanitizing robot

 Andrea Borgese, Dario Calogero Guastella, Giovanni Muscato, Stefania Stefani, Stefano Stracquadanio and Giuseppe Sutera
ABSTRACT. In this work, a robotic solution for UV-based disinfection and its efficacy evaluation are presented. Experimental trials have been performed to compare the virucidal efficacy of common omnidirectional fluorescent UV lamps with respect to our solution which leverages a mobile manipulator equipped with a UV source as an end-effector. Such a source can be thus moved and oriented in order to reach those surfaces which could be potentially hidden by other objects during the sanitizing process. This is achieved through a surface-aware 3D coverage path planning.

SA P2 2

Haptic Portable Pad for Hand Disease Manual Treatment

Mihai Dragusanu, Danilo Troisi, Alberto Villani, Domenico Prattichizzo and Monica Malvezzi
ABSTRACT. Nowadays, researchers are focusing their efforts on the remote distribution of devices for rehabilitation, allowing users to perform the exercises without the presence of a doctor. Physiotherapists, usually, employ manual therapy to treat patients with musculoskeletal discomfort and/or impairments. The purpose of this study is to introduce HAPP, a novel portable haptic device aimed to assist people suffering from various hand diseases, and more generally, to replicate the traditional mechanical and rhythmic stimulus features of manual treatments in order to study the effects of manual therapy in case of carpus and metacarpus problems. Its structure is composed of a mobile plate, actuated by a 3DOF parallel robot, with a rack-and-pinion mechanism that activates the end-effector to stimulate the user’s hand palm.

SA P2 3

A Concept for a Gravity-Balanced Upper-Limb Exoskeleton

Greta Vazzoler, Giovanni Berselli and Antonio Frisoli
ABSTRACT. This abstract reports a novel concept for a passive upper limb exoskeleton with 6 degrees of freedom. The device is conceived to support workers in industrial environments in a vast range of repetitive tasks. In this work, leveraging on a detailed analytical model developed in previous research, the best springs configuration to balance the system during the motion is designed through an efficient optimization routine. The model is validated by means of a multi-body tool for specific overhead tasks, and aspects concerning the embodiment design of the proposed balancer are evaluated.

SA P2 4

Conceptual Design of a Compliant, Low-Cost Prosthetic Hand

Mario Baggetta, Margherita Vazzoler, Giovanni Berselli, Gianluca Palli and Claudio Melchiorri
ABSTRACT. This paper addresses the conceptual design and the virtual prototyping of a low-cost, under-actuated prosthetic hand. Particular attention has been paid to ensuring that size, weight, and appearance are as similar as possible to those of a human hand, in order to increase user acceptance. The proposed device comprises a total of 10 Degrees of Freedom (DoF) and 4 Degrees of Actuation (DoA), three of which are enabled by employing electric motors and a tendon transmission system, whereas the last DoA is simply manually actuated by the end-user (e.g. using the still-functioning limb). The virtual prototype of the hand allows easy evaluation of the hand’s grasping capabilities as well as its maximum gripping force.

SA P2 5

Improving Surgical Robotics Training with Haptic Virtual Fixtures: An Experimental Study

Alberto Rota, Ke Fan and Elena De Momi
ABSTRACT. The lack of high-level assistive control strategies in the field of teleoperated surgical robots has been linked to intra-operative injuries and increased fatigue experienced by practitioners. Virtual Fixtures analogous to the one proposed in this study may be beneficial for the patient’s safety and the outcome of the operation; this work aims at evaluating their effectiveness in the context of surgical training. Tracking the position and orientation of a teleoperated surgical instrument with respect to a reference trajectory – planned in the pre-operative phase – allows one to compute and apply feedback forces to the manipulators held by the practitioner, which will provide haptic guidance towards an improved surgical performance. This high-level control strategy is here tested on a suturing task emulated in a virtual environment, where a group of participants was evaluated on the distance and angle error committed during the execution with and without assistance. The assistive modality proposed here is able to reduce the average error committed in the execution of the virtual suturing task: virtual fixtures and other similar assistance mechanisms may be most beneficial in the surgical training scenario, improving the learning curve and achieving better performances.

SA P2 6

Overview of a kinaesthetic hand exoskeleton system

Lorenzo Bartalucci, Nicola Secciani and Chiara Brogi
ABSTRACT. Within the new industrial era, the interaction between humans and virtual reality is spreading across our lives. The development of exoskeleton designed to enhance the immersivity of virtual reality environments has a potentially considerable social impact and arises as a hot research topic. The presented mechatronic design process of a kinaesthetic hand exoskeleton system meant to reproduce proprioceptive stimuli coming from the interaction with a virtual reality. The presented prototype is a modular device, equipped with force and pose sensors, and driven by a Bowden-cable-based remote actuation system. Unlike similar devices, the proposed exoskeleton is specifically thought for VR interaction and is designed to be reversible while exerting up to 15 N per finger.

SA P2 7

An optimized tactile sensing technology built for an anthropomorphic robotic hand

Avinash Kumar Singh, Petros Kaltsas and Fanny Ficuciello
ABSTRACT. The Prisma Hand II is an anthropomorphic multi functional robotic hand developed at PRISMA Lab, University of Naples Federico II which provides a solution for in-hand manipulation during grasping tasks. Each fingertip integrates a tactile/force sensor based on optoelectronic technology, providing tactile/force feedback during grasping and manipulation, particularly useful with deformable objects. The abstract proposes a new solution for the design and calibration of optimized sensors taking reference from the old sensors built based on the same technology.

SA P2 8

Actuation and Control of a Steerable Catheter for Mitral Valve Repair

Mattia Magro, Andrea Fortuna, Mariagrazia Quacquarelli, Angela Peloso, Xiu Zhang and Elena De Momi
ABSTRACT. In the field of Structural Heart Diseases, Mitral Regurgitation’s incidence is rising because of an aging population worldwide, and it has reached an annual mortality rate near 34%. The procedures of Structural Intervention Cardiology have enlarged the number of treated patients, since their minimally invasive and trans-catheter approach. To provide a forward step-change in this procedure, the aim of this work is to improve the use of the commercially available MitraClip system®, suggesting an innovative robot-assisted platform with autonomous control for the aforementioned system. The presented methodology is constituted of two phases: in the first one, we design, in the Solidworks® environment, 3D print and integrate the mechanical support with electrical motors and micro-controller devoted to catheter’s steering. In the second phase, we develop the closed-loop position control to improve the accuracy in the autonomous positioning of the catheter. The described approach was tested to demonstrate its feasibility and dexterity: a position accuracy of 1.1±0.54 mm in following a given optimal trajectory was obtained.

SA P2 9

Exploiting Robot Redundancy for Online Learning and Control

Marco Ficorilli, Marco Capotondi, Valerio Modugno and Alessandro De Luca
ABSTRACT. Accurate trajectory tracking in the task space is crit- ical in many robotics applications. Model-based robot controllers are able to ensure very good tracking but lose effectiveness in the presence of model uncertainties. On the other hand, online learning-based control laws can handle poor dynamic modeling, as long as prediction errors are kept small and decrease over time. However, in the case of redundant robots directly controlled in the task space, this condition is not usually met. We present an online learning-based control framework that exploits robot redundancy so as to increase the overall performance and shorten the learning transient. The validity of the proposed approach is shown through a comparative study conducted in simulation on a KUKA LWR4+ robot.

SA P2 10

RNN aided lidar-based positioning for AMR

Stefano Mutti and Nicola Pedrocchi
ABSTRACT. Autonomous Mobile Robots (AMR) are being employed more in nowadays industry, in tasks where a precise positioning algorithm is fundamental for a positive result. AMRs often rely on laser scanner readings in order to perform a precise positioning or docking procedure. Laser scanners output a 1D vector of distances between the sensor center and the surrounding environment. In order to re-position the AMR into a saved position, the AMR has to be moved in a way to match the sensor reading of the saved position with the current one. In this work, an RNN approach is proposed to estimate the registration error between laser scanners and improve the docking precision.

SA P2 11

Development of a Bi-level TO-based Co-design Pipeline for the RePAIR Robotic Platform

Andrea Patrizi, Arturo Laurenzi, Francesco Ruscelli, Eamon Barrett and Nikolaos Tsagarakis
ABSTRACT. We present a bi-level co-design pipeline employed for the optimization of a set of relevant relevant kinematic parameters of the bi-manual robotic platform under development for the european project RePAIR. In particular, the bottom level consists of a monolithic co-design framework, based on trajectory optimization, which jointly optimizes for design and state variables. The framework employs a kinematic model of the system and accounts for collisions, state bounds, a number of user-defined tasks, while minimizing a suitable performance index. The top level employs a three-step heuristic globalization algorithm which performs several calls to the low level TO to cope with the observed sensitivity of the TO solution w.r.t. the choice of the initial guess.

SA P2 12

Perception-aware trajectory planning for a pair of vehicles avoiding indistinguishability

Francesco Riz, Daniele Fontanelli and Luigi Palopoli
ABSTRACT. We consider a pair of unicycle vehicles, completely unaware of their position and orientation in the environment. By moving across the environment, each robot can rely on bounded range ranging measurements collected from a fixed-frame anchor and from the other vehicle. We propose a trajectory planning algorithm that allows the two robots to simultaneously reconstruct their state, i.e. to localise themselves in the environment.

SA P2 13

Fast Convex Visual Foothold Adaptation for Quadrupedal Locomotion

Shafeef Omar, Lorenzo Amatucci, Giulio Turrisi, Victor Barasuol and Claudio Semini
ABSTRACT. This extended abstract provides a short introduction on our recently developed perception-based controller for quadrupedal locomotion. Compared to our previous approach based on Visual Foothold Adaptation (VFA) and Model Predictive Control (MPC) \cite{b1}, our new framework combines a fast approximation of the safe foothold regions based on Neural Network regression, followed by a convex decomposition routine in order to generate safe landing areas where the controller can freely optimize the footholds location. The aforementioned framework, which combines prediction, convex decomposition, and MPC solution, is tested in simulation on our 140kg hydraulic quadruped robot (HyQReal).

SA P2 14

Efficient 2D LIDAR-Based Map Updating For Long-Term Operations in Dynamic Environments

Elisa Stefanini, Enrico Ciancolini, Alessandro Settimi and Lucia Pallottino
ABSTRACT. Long-time operations of autonomous vehicles and mobile robots in logistics and service applications are still a challenge. To avoid a continuous re-mapping, the map can be updated to obtain a consistent representation of the current environment. In this paper, we propose a novel LIDAR based occupancy grid map updating algorithm for dynamic environments. The proposed approach allows robust long-term operations as it can detect changes in the working area even in presence of moving elements. Results highlighting map quality and localisation performance, both in simulation and experiments, are reported.

SA P2 15

To Enabling Plant-like Movement Capabilities in Continuum Arms

Enrico Donato, Yasmin Tauqeer Ansari, Cecilia Laschi and Egidio Falotico
ABSTRACT. Enabling reaching capabilities in highly redundant continuum soft arms is an active area of research. So far, it has been heavily addressed through the brain-inspired notion of internal models, where sensory-motor spaces are correlated through learning-based computational frameworks. However, this work investigates an innovative source of bio-inspiration, i.e., plants, which can interestingly move towards a desired external stimulus despite the lack of a central nervous system, thereby, opening avenues to the development of a new generation of distributed control strategies for continuum arms. In particular, reaching is achieved through a combination of distributed sensing and curvature regulation. This work is a first translation of moving-by-growing mechanisms in plants intended to endow continuum and soft robotic arms with a novel repertoire of motions that can be exploited to efficiently navigate highly unstructured environments.

SA P2 16

Human reaction modeling for collaborative robotic applications in virtual environment

Adriano Scibilia
ABSTRACT. Model-based approaches aiming to characterize human behavior when interacting with a controlled machine have been a matter of research investigation in various domains, from aerospace to semi-autonomous driving and robotics. Human-robot collaboration is one of the most exciting scenarios of application in which a continuous physical interaction between humans and the controlled plant is present. In this context, the human subject can adapt its control behavior to the external sensed dynamics. This capability has a significant observable effect on the control delay, making its characterization and prevision a crucial aspect to understand. This work investigates a linear modeling approach that uniquely describes human and robot control actions and applies to a collaborative robotic task.

SA P2 17

Integrating Digital Twin And Mixed Reality In Human-Robot Collaboration

Mohammad Shaaban, Simone Macciò, Alessandro Carfì and Fulvio Mastrogiovanni
ABSTRACT. This paper proposes a novel software architecture for generalized scenarios of human-robot collaboration, in which a Digital Twin mirrors, monitors, and guides the interaction in real time, while Mixed Reality is used as a medium to ensure intuitive communication between agents involved.

SA P2 18

Regulation by Iterative Learning in Continuum Soft Robots

Marco Montagna, Pietro Pustina and Alessandro De Luca
ABSTRACT. The dynamic uncertainties and disturbances characterizing continuum soft robots call for the derivation of simple and possibly information-free controllers. We propose an iterative learning control law for shape regulation of continuum soft robots consisting of a PD action and a feedforward term, updated to learn the potential forces at the target configuration. We prove that the regulator achieves global asymptotic stabilization of the closed-loop system to the desired set-point. Simulation results validate the proposed control law.

SA P2 19

Towards autonomous soft grasping of deformable objects using flexible thin-film electro-adhesive gripper and online capacitance measure

Salvatore D’Avella, Ion-Dan Sîrbu, Marco Fontana, Rocco Vertechy and Paolo Tripicchio
ABSTRACT. Grasping fragile or deformable objects is a more complex task with respect to the traditional pick and place of solid objects. In such cases, a retention action is typically preferred over a compression force in order to avoid damaging the objects. The proposed work presents a robotic manipulation grasping system that leverages a gripper realized with the flexible thin-film electro-adhesive (EA) devices technology and a vision pipeline based on an RGB-D camera to detect the grasp pose configuration and track the target during the holding phase to check whether the task has been successfully completed. Several tests have been done to assess the capabilities of the proposed robotic system, picking and placing deformable objects, comparing the EA gripper with a traditional parallel jaw gripper. Furthermore, a self-sensing circuit capacitance has been developed for measuring the variation of the capacitance between electrodes of an EA device during the adhesion providing useful information to automatically detect the successful grip of an object and the possible loss of adhesion during manipulation.

SA P2 20

Task-oriented programming for industry: a comparison with robot-oriented programming

Michele Delledonne, Enrico Villagrossi, Marco Faroni, Manuel Beschi and Nicola Pedrocchi
ABSTRACT. The ease of use of robot programming interfaces represents a barrier to robot adoption in several manufacturing sectors because of the lack of expertise of the end-users. Current robot programming methods are mostly the past heritage, with robot programmers reluctant to adopt new programming paradigms. This work aims to evaluate the impact on non-expert users of introducing a new task-oriented programming interface that hides the complexity of a programming framework based on ROS. The paper compares the programming performance of such an interface with a classic robot-oriented programming method based on a state-of-the-art robot teach pendant. An experimental campaign involved 22 non-expert users working on the programming of two industrial tasks demonstrating a high acceptance level of the task-oriented interface with not significant difference in the learning time compared to a standard interface.

SA P2 21

RUVIFIST: Reconfigurable Underwater Vehicle for Inspection, Free-floating Intervention and Survey Tasks

Edoardo Topini, Gherardo Liverani, Jonathan Gelli, Cosimo Fredducci, Alessandro Ridolfi and Benedetto Allotta
ABSTRACT. The development of Autonomous Underwater Reconfigurable Vehicles (AURVs) with the capability of interacting with the surrounding environment and autonomously changing the configuration, according to the task at hand, can represent a real breakthrough in underwater system technologies. Driven by these considerations, an innovative AURV has been developed by the Department of Industrial Engineering of the University of Florence (DIEF), Italy, capable of efficiently reconfiguring its shape according to the task at hand. In particular, the RUVIFIST (Reconfigurable Underwater Vehicle for Inspection, Free-floating Intervention and Survey Tasks) vehicle has been provided with two extreme configurations: a slender (“survey”) configuration for long navigation tasks, and a stocky (“hovering”) configuration designed for challenging goals as intervention operations. Moreover, an accurate description of the overall vehicle is currently provided in this work, starting from the hardware architecture to the developed software framework.

SA P2 22

Mechatronic design of an underwater multisensor system for optical data acquisition

Alessandro Bucci, Cosimo Fredducci, Gherardo Liverani, Jonathan Gelli, Lorenzo Bartalucci, Francesco Ruscio, Vincenzo Manzari, Mirko Stifani, Riccardo Costanzi and Alessandro Ridolfi
ABSTRACT. The electromechanical design of a vision system, developed by the ISME nodes of the University of Florence (UNIFI) and the University of Pisa (UNIPI), Italy, is presented. The system is designed to test Visual Odometry algorithms and can be employed as a standalone device or in conjunction with an Autonomous Underwater Vehicle (AUV). In this regard, the module presents a wide variety of sensors and is intended to have its power supply. The project constraints are described, and the main features of the prototype are reported.

SA P2 23

Towards Culture-Aware Intelligent Systems

Ariel Gjaci, Luca Oneto, Carmine Tommaso Recchiuto and Antonio Sgorbissa
ABSTRACT. Understanding how the data used to train intelligent systems affects their behavior is a critical task in the artificial intelligence field. It is also known that making robots capable to adapt their actions according to the culture improves their interaction with humans. For this reason, it may be crucial to know how the cultural component inside data affects the prediction of intelligent systems. In this paper, we propose a method to acknowledge the cultural factor inside data and we show some preliminary results obtained by using a Random Forest model on two publicly available Datasets.

SA P2 24

Task Aware Impedance Planning

Liana Bertoni, Luca Muratore, Arturo Laurenzi and Nikos Tsagarakis
ABSTRACT. Robots in realistic environments require the capability of adapting themselves to different conditions and task requirements regarding motion tracking performances, interaction forces, and payload conditions. Therefore, an online impedance modulation during the task execution represents a critical prerequisite for the robots operating in real-world environments, remaining an open research topic today. This work presents a novel task-based online impedance modulation with stability online verified. The method allows modulating the robot impedance following the task requirements. The method considers the motion tracking performance requested by the task and the expected task payload or interaction forces to derive the robot impedance. Appropriate impedance levels are also maintained to ensure the stability of the modulated robot impedance. The method is successfully verified on the CENTAURO robot.



8 Ottobre, ore 15:30 – 18:30

Istituto Superiore Antincendi – Sala Caravaggio

SA P3 1

Towards Computer-Vision Based Vineyard Navigation for Quadruped Robots

Lee Milburn, Juan Gamba and Claudio Semini
ABSTRACT. There is a dramatic shortage of skilled labor for modern vineyards. The Vinum project is developing a mobile robotic solution to autonomously navigate through vineyards for winter grapevine pruning. This necessitates an autonomous navi- gation stack for the robot pruning a vineyard. The Vinum project is using the quadruped robot HyQReal. This paper introduces an architecture for a quadruped robot to autonomously move through a vineyard by identifying and approaching grapevines for pruning. The higher level control is a state machine switch- ing between searching for destination positions, autonomously navigating towards those locations, and stopping for the robot to complete a task. The destination points are determined by identifying grapevine trunks using instance segmentation from a Mask Region-Based Convolutional Neural Network (Mask- RCNN). These detections are sent through a filter to avoid redundancy and remove noisy detections. The combination of these features is the basis for the proposed architecture.

SA P3 2

Wearable Haptics for Interaction in Immersive VR in Children Neurorehabilitation

Daniele Leonardis, Cristian Camardella, Domenico Chiaradia and Antonio Frisoli
ABSTRACT. Immersive Virtual Reality has been investigated as an alternative and promising scenario to propose rehabilitation exercises to patients, with advantages in terms of flexibility and parametrization of the exercises, repeatability, and engagement. Modern, immersive VR systems provide novel features that are precious for usability of rehabilitation serious games in the clinical practice, envisaging also home-care applications. Still, the most of these impressive visual and audio virtual settings lack of the sense of touch: considering neurorehabilitation motor tasks, involving manipulation, the presence of haptic feedback is a key feature. We present here how haptic feedback can be introduced within an immersive serious game scenario, developed for rehabilitation of children with cerebral palsy. In the development, we considered different aspects including design of novel, highly wearable haptic devices, planning of the virtual rehabilitation exercise, and development of a dedicated haptic rendering strategy to provide an informative feedback during task execution.

SA P3 3

Design and Characterization of Modular Soft Exoskeleton for Hand Rehabilitation

Tommaso Bagneschi, Daniele Leonardis, Domenico Chiaradia and Antonio Frisoli
ABSTRACT. In this work, we present the investigation and the results of innovative soft finger modules of an actuated compact glove. The modular finger modules are based on a soft, open-ring structure, to improve the comfort of the user when the hand is relaxed, and at the same time to enhance the glove’s structural stability when it is active for grasping assistance. We present a novel modular design of the finger modules, integrated with a working prototype of the actuated glove. Design is then characterized and evaluated with experimental loading tests of the prototype parts and FEM simulation. The simulation helps to understand interface forces between the soft rings and the finger tissue which would be otherwise difficult to evaluate in experiments.

SA P3 4

GRACE: GeometRy-based Actuators that Contract and Elongate

Corrado De Pascali, Giovanna Adele Naselli, Stefano Palagi, Rob Bernardus Nicolaas Scharff and Barbara Mazzolai
ABSTRACT. This paper describes a recently developed class of 3D-printable soft pneumatic actuators, named GRACE. They can be built as a whole with the robotic artefact, avoiding structural discontinuities in the soft robot’s body. Their design is fully scalable and they can be built with different materials and additive manufacturing technologies. Their arrangements in series and/or parallel configurations enable the fabrication of complex bundles of muscle fibres for biomimetic machines.

SA P3 5

Towards Compact Design of Haptic Interfaces for Caress-Like Stimuli

Nicole D’Aurizio, Tommaso Lisini Baldi, Teresa Ramundo, Alessandro Moscatelli and Domenico Prattichizzo
ABSTRACT. Several studies in the affective haptics research field showed the potentiality of using haptic technology to convey emotions in remote communications. In this context, it is of interest to simplify the haptic feedback without altering the informative content of the stimulus to develop affective haptic devices whose technological complexity is limited and decrease the amount of data to be transmitted in remote haptic interactions. In this work, we investigated the correlation between the parameters regulating a caress-like stimulation and the perceived pleasantness. A small difference in the pleasantness ratings was observed between caresses provided with linear movements and those given as discrete sequences of taps. The presence of vibration had a little effect on the perceived pleasantness.

SA P3 6

A Hybrid Primitive-Based Planner for Autonomous Navigation with CENTAURO Robot

Alessio De Luca, Luca Muratore and Nikos Tsagarakis
ABSTRACT. Wheeled-legged robots have the ability to navigate in cluttered and irregular environments adapting the locomotion mode to the terrain perceived. To achieve this functionality, a locomotion planner is needed. In this work we present a hybrid search-based planner, which considers a set of modifiable motion primitives and a 2.5D traversability map acquired from the environment to generate navigation plans for the hybrid mobility robot CENTAURO. Our approach was validated in simulation and on the real wheeled-legged robot CENTAURO, demonstrating traversing capabilities in cluttered environments with various obstacles.

SA P3 7

An Information-aware Lyapunov-based MPC for autonomous robots

  Olga Napolitano, Daniele Fontanelli, Lucia Pallottino and Paolo Salaris
ABSTRACT. This paper proposes a feedback-feedforward control scheme that combines the benefits of an online active sensing control strategy to maximize the information needed for correctly executing the desired task (the feedforward component), with a Lyapunov-based control strategy that guarantees an asymptotic convergence towards the task itself (the feedback component). To show the effectiveness of our method, we consider a unicycle equipped with onboard sensors that has to perform the classical path following task.

SA P3 8

Inverse Reinforcement Learning algorithm for intra-vascular and intra-cardiac catheter’s navigation in Minimally Invasive Surgery

ABSTRACT. Structural Intervention Cardiology (SIC) is a miniinvasive intervention with a catheter based approach for cardiac surgery. Although SIC procedures are becoming increasingly popular, procedures are not ergonomic and technically demanding and, at the same time, high precision and accuracy in reaching target locations inside the human body are necessary for the success these procedures. Thus, there is therefore a need to develop a robust path planner framework to improve the accuracy in target reaching while minimizing interaction with anatomical structures. In this work a pre-operative path-planning method able to guide the catheter from the peripheral access to the desired target position with the needed orientation is proposed. The method exploits an Inverse Reinforcement Learning algorithm based on a combination of Behavioral Cloning (BC) and Generative Adversarial Imitation Learning (GAIL). The method was in-silico tested performing 50 intra-vascular and 70 intra-cardiac paths where the ratio between attempts in which the catheter reaches the target and total number of attempts, computation time, the difference between desired pose and the reached one were considered as validation metrics. Results show that the proposed method computes optimal path enabling the catheter to reach the target with an average error in position below 2 mm in the intra-vascular phase and below 1 mm in position and 6° in orientation in the intra-cardiac phase

SA P3 9

On the somatotopic mapping of acknowledgment haptic feedback of a sensorimotor interface

Leonardo Franco, Simone Rossi, Domenico Prattichizzo and Gionata Salvietti
ABSTRACT. Supernumerary Robotic Limbs (SRL) represent a new class of robots whose aim is to augment the human body’s possibilities, for example extending its workspace. To control them the user must use a dedicated interface, which can both extract inputs from the human body and provide feedback about the robot’s status. Such a device is called a sensorimotor interface and the feedback subsystem is usually a wearable haptic device. The brain’s human somatotopic arrangement in the central nervous system lacks a location for artificially added limbs. In this extended abstract, we investigate if there is the best location for feedback coming from a robot not directly associated with a part of the wearer’s body. We have tested four different body locations – shoulder, wrist, hip, and ankle – for vibrotactile feedback coming from the simulated interaction with an extra robotic limb activated using an interface consisting of an accelerometer worn on the user’s shoulder. Results from the experiment involving 14 participants demonstrated that the ankle feedback position led to significantly worse performances when having input from the shoulder, whereas the other three locations led to comparable results.

SA P3 10

Design for Values (DfV) in Action. The case of Robotics for Social Sustainability

Marianna Capasso, Alberto Pirni, Gastone Ciuti and Paolo Dario
ABSTRACT. A plurality of methodologies and frameworks have been proposed in recent years to embed robotics and other emerging technologies in human social practices, in line with the policy-aims of Responsible Research and Innovation (RRI). In this paper, we discuss and explore recent findings related to Design for Values (DfV), which is a systematic and theoretically grounded approach that tries to integrate design processes with human social and moral values. Such findings will be analysed in light of specific robotic platforms, and by discussing and reviewing some of the results we recently published. The main objective is to show how and to what extent DfV may constitute a coherent methodology to incorporate dimensions of social sustainability into robotic design and development. This may serve to further advance the understanding of robotics as an inherently cross-domain field of studies, as primarily laid out in the scientific mission of Robot Companions (RCs), i.e., robots that will be designed to create new sustainable, affordable, and socially beneficial solutions for all citizens.

SA P3 11

Adaptive human tracking using wearable sensors

Ilaria Cetera, Marco Bardanzellu, Laura Giarrè and Federica Pascucci
ABSTRACT. In the last decade, the indoor localization and positioning has gained attention. The development of location based service, such as activity monitoring during training, has been one of the main drivers for this interest. In this contribution, a tracking system based on inertial measurement is proposed. The tracking system is designed for human tracking. The system has been designed to be low cost and self-configurable. To this aim, the system is able to learn all the parameters exploited in the computation. Moreover, the computational load of the system is low, so it can be implemented over wearable devices. The proposed approach has been tested in a real scenario and the results are reported.

SA P3 12

Unsupervised learning of quadruped robot balance

Paolo Arena, Fabio Di Pietro, Alessia Li Noce and Luca Patanè
ABSTRACT. This paper reports a research activity grounded on year-long experience in unsupervised neural networks used in a lot of different scenarios and here applied to control balance in a quadrupedal robot structure. Simulation results are reported on balance control of a mini cheetah quadruped robot undergoing unexpected weight disturbances coming from additional loads applied on the robot body. The unsupervised learning structure, belonging to the family of Motor Maps controllers, is shown to rapidly learn the additional torques to be applied to the robot legs in order to compensate for the disturbance and so contributing to improve the underlying Whole Body Impulse Controller in front of the robot model uncertainties.

SA P3 13

Considerations on possible approaches to measure risk for obstacle avoidance

Lorenzo Paiola, Giorgio Grioli and Antonio Bicchi
ABSTRACT. Risk minimization has historically been tackled with chance constraints or with risk-aware measure acting on stochastic cost functions. To characterize risk in an obstacle avoidance setting, the computation of the probability of collision is of paramount importance. This paper explores and compares two approaches to compute such probabilities for a robot and an obstacle under Gaussian uncertainty along a continuous path. We first establish a theoretical framework, show numerical simulations, and finally we highlight the advantages and shortcomings of the considered approaches.

SA P3 14

Autonomous underwater environment perceiving and modeling: an experimental campaign with FeelHippo AUV for Forward Looking Sonar-based Automatic Target Recognition and Data Association

Alberto Topini, Leonardo Zacchini, Lorenzo Bazzarello, Vincenzo Manzari, Nicola Secciani, Matteo Franchi, Mirko Stifani and Alessandro Ridolfi
ABSTRACT. Seabed inspection is one of the most sought-after applications for Autonomous Underwater Vehicles (AUVs). In particular, acoustic sensors, as Side Scan Sonars (SSSs) and Forward-Looking Sonars (FLSs), are commonly favored over optical cameras to carry out such a task being not influenced by the illumination conditions and providing high-range data. However, acoustic images are often hard to interpret with conventional automatic techniques, forcing human operators to analyze thousands of collected images to identify the so-called Objects of Potential Interest (OPIs). In this perspective, this paper reports the development of an Automatic Target Recognition (ATR) methodology to identify and localize OPIs in FLS imagery; such detections have been then exploited to realize a world model with the Probabilistic Multiple Hypothesis Anchoring (PMHA) data association algorithm. Both the ATR and world modeling systems were on-field tested at the Naval Support and Experimentation Centre (Centro di Supporto e Sperimentazione Navale – CSSN) basin, in La Spezia, Italy, in a multi-vehicle architecture by employing an acoustic channel between FeelHippo AUV and an autonomous moving buoy.

SA P3 15

3D functional sport prostheses

Michela Bogliolo, Lea Turolla, Maura Casadio, James Segre and Elena Paradoi
ABSTRACT. 3D printing techniques are making rapid improvements, especially in the medical field, where many people are requesting prosthetic devices that are customized in design and function. This is happening especially in sports to facilitate access for the disabled, providing more affordable devices suitable for those who want to play sports without having Olympic ambitions. In this study we present some 3D printed sport devices, for swimming and cycling, developed by the Io Do Una Mano Association, the official Italian chapter of e-Nable. Each device presented has been customized for specific disabilities and is based on different requests.

SA P3 16

Robot Kinesthetic Teaching Enhanced by sEMG-based Estimation of Muscle Co-Contraction and Bio-Feedback

Roberto Meattini, Davide Chiaravalli, Gianluca Palli and Claudio Melchiorri
ABSTRACT. sEMG signals are exploited for unsupervised estimation of the co-contraction level of forearm’s muscles. In this way, by also exploiting a feedback based on a vibrotactile bracelet, the ability of operators in stiffening their hand was evaluated during kinesthetic teaching, in order to regulate the estimated co-contraction level to (i) match reference levels and (ii) activate the opening/closing of a gripper, i.e. in using their myoelectric signals enhance robot kinesthetic teaching operations. Experiments were carried out. The results provide positive outcomes on the intuitiveness and effectiveness of the proposed system and approach.

SA P3 17

Finger actuation of a Modular Wearable Exoskeleton for Hand/Wrist Rehabilitation and Training

Mihai Dragusanu, Danilo Troisi, Tommaso Lisini Baldi, Zubair Iqbal, Domenico Prattichizzo and Monica Malvezzi
ABSTRACT. Robot rehabilitation is an emerging and promising topic that incorporates robotics with rehabilitation and neuroscience to define new methods for supporting patients with diseases. The rehabilitation process could increase the efficacy exploiting the potentialities of robot-mediated therapies. In this paper, we describe the hand module of a system for hand/wrist motion training. The whole system is designed to be wearable, easy to control and manage. It can be used by the patient in collaboration with the therapist or autonomously. The paper introduces the main steps of device design and development and presents some possible exercises that can be performed by a user with limited wrist mobility.

SA P3 18

Risk space modelling for human-robot collaboration in a shared intralogistics scenario

Elena Stracca, Giorgio Grioli, Lucia Pallottino and Paolo Salaris
ABSTRACT. This paper aims to define a multi-dimensional risk space for a robot moving in a shared environment with human operators. An analysis of the robot state variables on which each individual risk source may depend leads to the introduction of a fuzzy inference system to quantify the risk levels. The methodology presented in this paper outlines a general way to characterize risk for autonomous agents working in partially unknown environments.

SA P3 19

State Estimation for Agile Quadrotors in the Wild

Giovanni Cioffi and Davide Scaramuzza
ABSTRACT. In this extended abstract, we present our latest research in robust state estimation for agile quadrotor flight. We discuss the differences between discrete- and continuous- time trajectory representations in visual-inertial odometry (VIO). Then, we present a novel VIO algorithm that combines events and standard frames to estimate the pose of a quadrotor subject to a rotor failure. Finally, we discuss our recent progress in learned inertial odometry for quadrotor flight. We conclude with the next research directions that have the potential to improve the robustness of onboard state estimation systems for autonomous drones.

SA P3 20

A Framework for Multi-Robot Task and Motion Planning

Antonio Iannone, Antony Thomas and Fulvio Mastrogiovanni
ABSTRACT. We present an approach for multi-robot integrated task and motion planning that allocates tasks to robots. Specifically, we focus on transportation-like tasks where the objective of a task correspond to transporting objects to a desired location. We demonstrate the performance under varying robot number and transportation tasks.

SA P3 21

Robot Perception through Wearable Sensors: Decoding Grasping for Human-Robot Hand-Over

Andrea Bonci, Laura Burattini, Sandro Fioretti, Maria Cristina Giannini, Sauro Longhi, Alessandro Mengarelli, Andrea Tigrini and Federica Verdini
ABSTRACT. Human-robot interaction represents the cornerstone for the full development of Industry 4.0 and 5.0 paradigms, that rely on this cooperation in order to develop more efficient and flexible production lines. In this context, the human-robot handover plays a crucial role and many approaches were introduced to plan and control this task, including the less investigated decoding of human muscles activity. Hence, the design of reliable myoelectric human-robot interfaces is a point of primary interest. This paper investigates the use of a wearable device, i.e. an armband, for achieving a robust detection of several human grasping gestures. An evaluation of the most useful features, belonging to three different computational domains, is also proposed. Outcomes showed that high recognition performance can be achieved with limited computational burden, which is crucial when dealing with real-time demands in collaborative task.

SA P3 22

Output feedback control of a UAV for vision-based target tracking

Mirko Leomanni, Francesco Ferrante, Nicholas Cartocci, Gabriele Costante and Mario Luca Fravolini
ABSTRACT. In recent years, small unmanned aerial vehicles (UAVs) such as quadrotors have become a popular platform for control and robotics research with application to mapping, delivery, and surveillance problems, to name a few . The research activity in autonomous navigation has increased dramatically and the proposed solutions are rapidly approaching the goal of near-complete autonomy. To achieve this ambitious goal, it is also necessary to provide the vehicle with autonomous maneuvering capabili- ties. In this respect, the incorporation of visual cues in the flight control system is seen as a key element for designing robust and reactive control laws. In this paper, a model-based control approach is proposed for autonomous vision-based tracking of a moving target by a quadrotor vehicle. To solve this problem, we rely on an output feedback perspective within a robust control framework.

SA P3 23

Functional Iterative Learning Control for Linear Systems

Franco Angelini and Cosimo Della Santina
ABSTRACT. Many linear dynamic systems can usually be modeled in continuous time. However, to simplify the learning phase of Iterative Learning Control, the system output signal is frequently sampled, while the input is discretized, de-facto obtaining a discrete-time system. This work aims at preserving the continuous nature of the system, while tracking sampled outputs. To this end, we model the input signal as an element of a functional subspace, exploiting its infinite-dimension. Simulation results are presented.

SA P3 24

TelePhysicalOperation: a Shared Control Architecture for Intuitive and Smart Teleoperation of Complex Mobile Manipulators

Davide Torielli, Luca Muratore and Nikos Tsagarakis
ABSTRACT. Recent years have shown a growing demand toward automation in the industry and other contexts. To follow this trend, the research has to face some challenges to exploit the capabilities of complex robotic systems. Heavy redundant platforms, like dual arm manipulators, mobile robots, and legged systems, can help in accomplishing always more difficult tasks, but they also need more effort to operate with them. In this scenario, we have faced the challenge of exploring new teleoperation interfaces. With the development of the TelePhysicalOperation architecture, we want to provide: (1) an intuitive interface to permit even to a non expert user to control complex robots; and (2) more robot autonomy capabilities, to reduce the operator burden and to improve the performance of the teleoperation task. This paper recaps our recent work done in this context, including experimental validations with the CENTAURO robot, a dual arm platform equipped with a hybrid leg-wheel mobile system.