arXiv subjects

Claudio Pacchierotti

Publications and source records attributed to Claudio Pacchierotti.

13 recordsLinked to original sources

arXiv2026

Sampling-Based Coordination-Informed Multi-Objective Multi-Robot Reinforcement Learning

Multi-robot systems must simultaneously optimize competing objectives while maintaining coordinated behavior. Existing multi-agent reinforcement learning approaches often rely on fixed or centralized coordination, which limits adaptability and violates distributed constraints. This work introduces the Coordination-Informed Multi-Objective Reinforcement Learning (CIMORL) framework, integrating a distributed weight prediction mechanism, a privileged expert training strategy, and theoretical guarantees for Pareto-optimal solutions. We present the base CIMORL method alongside two sampling-based variants, CIMORL-TS (Tree Search) and CIMORL-MPPI (MPPI), which leverage privileged global information during training to enable fully decentralized deployment. Experimental validation in cooperative and adversarial scenarios demonstrates a $21.2\%$ hypervolume improvement and superior policy stability compared to state-of-the-art baselines. Real-world experiments with Crazyflie drones further validate the framework's robustness in resource allocation and multi-attacker multi-defend scenarios under partial observability.

cs.RO↗

arXiv2026

A Gated Graph Neural Network Approach to Fast-Convergent Dynamic Average Estimation

Dynamic average estimation is a critical problem in multi-agent systems, enabling agents to collaboratively estimate time-varying signals using only local information exchange. Traditional model-based approaches often face challenges related to convergence speed and sensitivity to network topology changes. This paper introduces a novel learning-based solution leveraging Gated Graph Neural Networks (GGNNs) for fast-convergent dynamic average estimation in a fully distributed manner. Taking advantage of the inherent structure of GGNNs, the proposed method models the estimation process as a distributed autoregressor, ensuring rapid convergence while maintaining stability. We incorporate a regularization term during training to enforce convergence guarantees and introduce an encoding-decoding mechanism to reduce communication overhead without sacrificing accuracy compared to standard GGNNs. Extensive numerical experiments demonstrate that our approach significantly outperforms conventional model-based estimators in terms of both convergence speed and precision, making it a promising alternative for multi-agent applications that require dynamic average estimation.

cs.LG↗

arXiv2025

Decentralized Reinforcement Learning for Multi-Agent Multi-Resource Allocation via Dynamic Cluster Agreements

This paper addresses the challenge of allocating heterogeneous resources among multiple agents in a decentralized manner. Our proposed method, Liquid-Graph-Time Clustering-IPPO, builds upon Independent Proximal Policy Optimization (IPPO) by integrating dynamic cluster consensus, a mechanism that allows agents to form and adapt local sub-teams based on resource demands. This decentralized coordination strategy reduces reliance on global information and enhances scalability. We evaluate LGTC-IPPO against standard multi-agent reinforcement learning baselines and a centralized expert solution across a range of team sizes and resource distributions. Experimental results demonstrate that LGTC-IPPO achieves more stable rewards, better coordination, and robust performance even as the number of agents or resource types increases. Additionally, we illustrate how dynamic clustering enables agents to reallocate resources efficiently also for scenarios with discharging resources.

stat.ML↗

arXiv2024

Evaluation of Human-Robot Interfaces based on 2D/3D Visual and Haptic Feedback for Aerial Manipulation

Most telemanipulation systems for aerial robots provide the operator with only 2D screen visual information. The lack of richer information about the robot's status and environment can limit human awareness and, in turn, task performance. While the pilot's experience can often compensate for this reduced flow of information, providing richer feedback is expected to reduce the cognitive workload and offer a more intuitive experience overall. This work aims to understand the significance of providing additional pieces of information during aerial telemanipulation, namely (i) 3D immersive visual feedback about the robot's surroundings through mixed reality (MR) and (ii) 3D haptic feedback about the robot interaction with the environment. To do so, we developed a human-robot interface able to provide this information. First, we demonstrate its potential in a real-world manipulation task requiring sub-centimeter-level accuracy. Then, we evaluate the individual effect of MR vision and haptic feedback on both dexterity and workload through a human subjects study involving a virtual block transportation task. Results show that both 3D MR vision and haptic feedback improve the operator's dexterity in the considered teleoperated aerial interaction tasks. Nevertheless, pilot experience remains the most significant factor.

cs.RO↗

arXiv2024

Multi-UAVs end-to-end Distributed Trajectory Generation over Point Cloud Data

This paper introduces an end-to-end trajectory planning algorithm tailored for multi-UAV systems that generates collision-free trajectories in environments populated with both static and dynamic obstacles, leveraging point cloud data. Our approach consists of a 2-fork neural network fed with sensing and localization data, able to communicate intermediate learned features among the agents. One network branch crafts an initial collision-free trajectory estimate, while the other devises a neural collision constraint for subsequent optimization, ensuring trajectory continuity and adherence to physicalactuation limits. Extensive simulations in challenging cluttered environments, involving up to 25 robots and 25% obstacle density, show a collision avoidance success rate in the range of 100 -- 85%. Finally, we introduce a saliency map computation method acting on the point cloud data, offering qualitative insights into our methodology.

cs.MA↗

arXiv2024

Liquid-Graph Time-Constant Network for Multi-Agent Systems Control

In this paper, we propose the Liquid-Graph Time-constant (LGTC) network, a continuous graph neural network(GNN) model for control of multi-agent systems based on therecent Liquid Time Constant (LTC) network. We analyse itsstability leveraging contraction analysis and propose a closed-form model that preserves the model contraction rate and doesnot require solving an ODE at each iteration. Compared todiscrete models like Graph Gated Neural Networks (GGNNs),the higher expressivity of the proposed model guaranteesremarkable performance while reducing the large amountof communicated variables normally required by GNNs. Weevaluate our model on a distributed multi-agent control casestudy (flocking) taking into account variable communicationrange and scalability under non-instantaneous communication

cs.MA↗

arXiv2024

Transferring human emotions to robot motions using Neural Policy Style Transfer

Neural Style Transfer (NST) was originally proposed to use feature extraction capabilities of Neural Networks as a way to perform Style Transfer with images. Pre-trained image classification architectures were selected for feature extraction, leading to new images showing the same content as the original but with a different style. In robotics, Style Transfer can be employed to transfer human motion styles to robot motions. The challenge lies in the lack of pre-trained classification architectures for robot motions that could be used for feature extraction. Neural Policy Style Transfer TD3 (NPST3) is proposed for the transfer of human motion styles to robot motions. This framework allows the same robot motion to be executed in different human-centered motion styles, such as in an angry, happy, calm, or sad fashion. The Twin Delayed Deep Deterministic Policy Gradient (TD3) network is introduced for the generation of control policies. An autoencoder network is in charge of feature extraction for the Style Transfer step. The Style Transfer step can be performed both offline and online: offline for the autonomous executions of human-style robot motions, and online for adapting at runtime the style of e.g., a teleoperated robot. The framework is tested using two different robotic platforms: a robotic manipulator designed for telemanipulation tasks, and a humanoid robot designed for social interaction. The proposed approach was evaluated for both platforms, performing a total of 147 questionnaires asking human subjects to recognize the human motion style transferred to the robot motion for a predefined set of actions.

cs.RO↗

arXiv2024

Neural Style Transfer with Twin-Delayed DDPG for Shared Control of Robotic Manipulators

Neural Style Transfer (NST) refers to a class of algorithms able to manipulate an element, most often images, to adopt the appearance or style of another one. Each element is defined as a combination of Content and Style: the Content can be conceptually defined as the what and the Style as the how of said element. In this context, we propose a custom NST framework for transferring a set of styles to the motion of a robotic manipulator, e.g., the same robotic task can be carried out in an angry, happy, calm, or sad way. An autoencoder architecture extracts and defines the Content and the Style of the target robot motions. A Twin Delayed Deep Deterministic Policy Gradient (TD3) network generates the robot control policy using the loss defined by the autoencoder. The proposed Neural Policy Style Transfer TD3 (NPST3) alters the robot motion by introducing the trained style. Such an approach can be implemented either offline, for carrying out autonomous robot motions in dynamic environments, or online, for adapting at runtime the style of a teleoperated robot. The considered styles can be learned online from human demonstrations. We carried out an evaluation with human subjects enrolling 73 volunteers, asking them to recognize the style behind some representative robotic motions. Results show a good recognition rate, proving that it is possible to convey different styles to a robot using this approach.

cs.RO↗

arXiv2023

Input State Stability of Gated Graph Neural Networks

In this paper, we aim to find the conditions for input-state stability (ISS) and incremental input-state stability ($\delta$ISS) of Gated Graph Neural Networks (GGNNs). We show that this recurrent version of Graph Neural Networks (GNNs) can be expressed as a dynamical distributed system and, as a consequence, can be analysed using model-based techniques to assess its stability and robustness properties. Then, the stability criteria found can be exploited as constraints during the training process to enforce the internal stability of the neural network. Two distributed control examples, flocking and multi-robot motion control, show that using these conditions increases the performance and robustness of the gated GNNs.

cs.RO↗

arXiv2022

Action-Specific Perception & Performance on a Fitts's Law Task in Virtual Reality: The Role of Haptic Feedback

While user's perception & performance are predominantly examined independently in virtual reality, the Action-Specific Perception (ASP) theory postulates that the performance of an individual on a task modulates this individual's spatial & time perception pertinent to the task's components & procedures. This paper examines the association between performance & perception & the potential effects that tactile feedback modalities could generate. This paper reports a user study (N=24), in which participants performed a Fitts's law target acquisition task by using three feedback modalities: visual, visuo-electrotactile, & visuo-vibrotactile. The users completed 3 Target Sizes X 2 Distances X 3 feedback modalities = 18 trials. The size perception, distance perception, & (movement) time perception were assessed at the end of each trial. Performance-wise, the results showed that electrotactile feedback facilitates a significantly better accuracy compared to vibrotactile & visual feedback, while vibrotactile provided the worst accuracy. Electrotactile & visual feedback enabled a comparable reaction time, while the vibrotactile offered a substantially slower reaction time than visual feedback. Although amongst feedback types the pattern of differences in perceptual aspects were comparable to performance differences, none of them was statistically significant. However, performance indeed modulated perception. Significant action-specific effects on spatial & time perception were detected. Changes in accuracy modulate both size perception & time perception, while changes in movement speed modulate distance perception. Also, the index of difficulty was found to modulate perception. These outcomes highlighted the importance of haptic feedback on performance, & importantly the significance of action-specific effects on spatial & time perception in VR, which should be considered in future VR studies.

cs.HC↗

arXiv2021

Electrotactile Feedback For Enhancing Contact Information in Virtual Reality

This paper presents a wearable electrotactile feedback system to enhance contact information for mid-air interactions with virtual objects. In particular, we propose the use of electrotactile feedback to render the interpenetration distance between the user's finger and the virtual content is touched. Our approach consists of modulating the perceived intensity (frequency and pulse width modulation) of the electrotactile stimuli according to the registered interpenetration distance. In a user study (N=21), we assessed the performance of four different interpenetration feedback approaches: electrotactile-only, visual-only, electrotactile and visual, and no interpenetration feedback. First, the results showed that contact precision and accuracy were significantly improved when using interpenetration feedback. Second, and more interestingly, there were no significant differences between visual and electrotactile feedback when the calibration was optimized and the user was familiarized with electrotactile feedback. Taken together, these results suggest that electrotactile feedback could be an efficient replacement of visual feedback for enhancing contact information in virtual reality avoiding the need of active visual focus and the rendering of additional visual artefacts.

cs.HC↗

arXiv2021

Electrotactile feedback applications for hand and arm interactions: A systematic review, meta-analysis, and future directions

Haptic feedback is critical in a broad range of human-machine/computer-interaction applications. However, the high cost and low portability/wearability of haptic devices remain unresolved issues, severely limiting the adoption of this otherwise promising technology. Electrotactile interfaces have the advantage of being more portable and wearable due to their reduced actuators' size, as well as their lower power consumption and manufacturing cost. The applications of electrotactile feedback have been explored in human-computer interaction and human-machine-interaction for facilitating hand-based interactions in applications such as prosthetics, virtual reality, robotic teleoperation, surface haptics, portable devices, and rehabilitation. This paper presents a technological overview of electrotactile feedback, as well a systematic review and meta-analysis of its applications for hand-based interactions. We discuss the different electrotactile systems according to the type of application. We also discuss over a quantitative congregation of the findings, to offer a high-level overview into the state-of-art and suggest future directions. Electrotactile feedback systems showed increased portability/wearability, and they were successful in rendering and/or augmenting most tactile sensations, eliciting perceptual processes, and improving performance in many scenarios. However, knowledge gaps (e.g., embodiment), technical (e.g., recurrent calibration, electrodes' durability) and methodological (e.g., sample size) drawbacks were detected, which should be addressed in future studies.

cs.HC↗

arXiv2012

Cutaneous Force Feedback as a Sensory Subtraction Technique in Haptics

A novel sensory substitution technique is presented. Kinesthetic and cutaneous force feedback are substituted by cutaneous feedback (CF) only, provided by two wearable devices able to apply forces to the index finger and the thumb, while holding a handle during a teleoperation task. The force pattern, fed back to the user while using the cutaneous devices, is similar, in terms of intensity and area of application, to the cutaneous force pattern applied to the finger pad while interacting with a haptic device providing both cutaneous and kinesthetic force feedback. The pattern generated using the cutaneous devices can be thought as a subtraction between the complete haptic feedback (HF) and the kinesthetic part of it. For this reason, we refer to this approach as sensory subtraction instead of sensory substitution. A needle insertion scenario is considered to validate the approach. The haptic device is connected to a virtual environment simulating a needle insertion task. Experiments show that the perception of inserting a needle using the cutaneous-only force feedback is nearly indistinguishable from the one felt by the user while using both cutaneous and kinesthetic feedback. As most of the sensory substitution approaches, the proposed sensory subtraction technique also has the advantage of not suffering from stability issues of teleoperation systems due, for instance, to communication delays. Moreover, experiments show that the sensory subtraction technique outperforms sensory substitution with more conventional visual feedback (VF).

cs.RO↗