Paolo Robuffo Giordano

Antonio Marino, Esteban Restrepo, Soon-jo Chung et al.

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.

Raul Fernandez-Fernandez, Marco Aggravi, Paolo Robuffo Giordano et al.

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.

Antonio Marino, Esteban Restrepo, Claudio Pacchierotti et al.

This paper addresses the challenge of allocating heterogeneous resources among multiple agents in a decentralized manner. Our proposed method, LGTC-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.

Amir M. Ghalamzan Esfahani, Firas Abi-Farraj, Paolo Robuffo Giordano et al.

This paper addresses the problem of mixed initiative, shared control for master-slave grasping and manipulation. We propose a novel system, in which an autonomous agent assists a human in teleoperating a remote slave arm/gripper, using a haptic master device. Our system is designed to exploit the human operator's expertise in selecting stable grasps (still an open research topic in autonomous robotics). Meanwhile, a-priori knowledge of: i) the slave robot kinematics, and ii) the desired post-grasp manipulative trajectory, are fed to an autonomous agent which transmits force cues to the human, to encourage maximally manipulable grasp pose selections. Specifically, the autonomous agent provides force cues to the human, during the reach-to-grasp phase, which encourage the human to select grasp poses which maximise manipulation capability during the post-grasp object manipulation phase. We introduce a task-relevant velocity manipulability cost function (TOV), which is used to identify the maximum kinematic capability of a manipulator during post-grasp motions, and feed this back as force cues to the human during the pre-grasp phase. We show that grasps which minimise TOV result in significantly reduced control effort of the manipulator, compared to other feasible grasps. We demonstrate the effectiveness of our approach by experiments with both real and simulated robots.