arXiv subjects

Ruggero Carli

Publications and source records attributed to Ruggero Carli.

At least 19 recordsLinked to original sources

arXiv2026

Submodular Policy Learning for Distributed Task Allocation in Open Multi-Agent Systems

This paper studies policy learning for distributed task allocation in open multi-agent systems, where agents may join and leave in a time-varying fashion, with submodular stage team utilities. At each time, the active agents select actions from local categorical policies such that the feasible joint agent-action pairs form a partition matroid. Standard continuous relaxations of submodular set functions are based on independent Bernoulli sampling, making them inconsistent with agents' policies.To solve this mismatch, we propose the \emph{partition multilinear extension} (PME), a policy-based relaxation whose continuous support matches feasible actions under categorical policies.We prove that the marginal gains of the stage utility provide an unbiased estimator of the gradient of the PME and that maximizing the PME over action distributions is equivalent to maximizing the stage utilities over agent actions, which are critical to devise principled policy gradient.Building on this, we design \emph{SubMAPL}, a centralized-training decentralized-execution KL-mirror policy-learning method that uses local marginal gains as stochastic PME gradients during training. KL-mirror updates preserve categorical feasibility without Euclidean projection.In the case where agents run tabular-softmax policies, we introduce open policy migration and an open-system KL tracking variation to handle agent arrivals and departures. Using dynamic regret analysis, we establish a lower bound on the cumulative utility which accounts for the openness of the environment and for the gap between optimal stage-wise and global utilities. Simulations on multi-agent coverage demonstrate that SubMAPL outperforms policy-gradient and online-learning baselines.

cs.MA↗

arXiv2026

Hierarchical Topology-Aware Planning and Control of Underwater Vehicle-Manipulator Systems in Confined Environments

This paper addresses autonomous intervention with an underwater vehicle--manipulator system (UVMS) in confined, cluttered, and partially known environments, where poor maneuverability, narrow passages, and uncertain execution may cause the robot to enter unrecoverable regions. We propose MANTA, a three-layer hierarchical planning-and-control framework that couples passage accessibility, manipulation feasibility, and closed-loop execution. The first layer performs global connectivity reasoning in a conservative reduced base space to extract traversable corridor candidates toward the task region. The second layer refines each candidate corridor by jointly optimizing the continuous base motion and arm trajectory, producing a collision-free base--arm trajectory. The third layer learns a reach-and-hold base policy using Gaussian-process model-based reinforcement learning (MBRL) through MC-PILCO, enabling trajectory tracking and station keeping at the planned manipulation state. During execution, the framework monitors map updates and can trigger recovery and route repair when the active passage becomes infeasible. MANTA is evaluated in confined UVMS planning and closed-loop tracking experiments. Across 120 matched planning queries, it achieves higher task success than full-state sampling-based baselines while producing larger clearance margins and lower arm motion. The learned MC-PILCO policy further reduces position and yaw tracking errors on both training and unseen tube-like references. These results show MANTA as a structured and data-efficient framework for safe autonomous underwater intervention in caves, tubes, and cluttered subsea structures.

cs.RO↗

arXiv2026

Physics-Informed Condition Monitoring of SiC Power Modules

Silicon carbide (SiC) power modules are increasingly deployed in automotive traction inverters, where condition monitoring is essential to prevent in-service failures. Despite extensive qualification under AQG 324, no consolidated approach exists for in-field health state estimation: physics-of-failure lifetime models lack real-time applicability, purely data-driven architectures require large labeled datasets and generalize poorly, and physics-informed frameworks remain too demanding for embedded deployment. We address SiC MOSFET modules assembled with sintered packaging, which suppresses solder degradation and produces aging behavior distinct from previously studied devices. Instead of the smooth quasi-exponential drift of solder-based modules, the forward voltage drop $V_{DS}$ exhibits multi-regime profiles, with wirebond liftoff events introducing abrupt, non-monotonic perturbations. We propose a condition monitoring framework combining three elements. First, physics-informed features replace raw sensor signals with cumulative damage indicators derived from junction temperature swing, mean junction temperature and a Miner rule accumulator, encoding degradation history in an interpretable form. Second, a monotonicity constraint enforced by gradient penalty regularization embeds the expected degradation direction as a physics-guided prior. Third, a heavy-tailed output distribution replaces the point estimate, giving calibrated uncertainty robust to the out-of-distribution variance introduced by liftoff. On an industrial power cycling dataset from Infineon Technologies, several neural architectures are compared under a strict cross-validation protocol. The full configuration reduces mean absolute error by approximately 70% over purely data-driven baselines and stays stable across all folds, while remaining lightweight enough for embedded deployment.

eess.SY↗

arXiv2026

Failure-Mechanism Transferability of Cumulative-Damage Features for Health State Estimation of SiC Power Modules

Data-driven health-state estimators for SiC (Silica-Carbide) power modules typically report their performance on a single accelerated-aging campaign, and how that performance transfers to a different failure mechanism is rarely tested. We benchmark five reference methods from the prognostics and condition-monitoring literature against a physics-informed NODE (Neural Ordinary Differential Equation) on two SiC power-cycling campaigns driven by structurally different failure mechanisms, solder-layer fatigue and wire-bond lift-off, under a per-module $k$-fold protocol. The NODE is evaluated under two input regimes that share the rest of the pipeline: the baseline electrical precursors and a set of cumulative thermoelectric features. Every reference method degrades on the wire-bond campaign, with average errors growing and precision decreasing with respect to their performance on the soldered campaign. The NODE fed with the cumulative features keeps its soldered-campaign metrics on both mechanisms, with differences inside the fold-to-fold variance, while the same architecture fed with the baseline precursors falls back to the reference-method cluster. The input representation contributes at least as much as the architecture to failure-mechanism transferability of a health-state estimator.

eess.SY↗

arXiv2026

Pursuing Optimal Stepsize in Adaptive Gradient-Based Quadratic Optimization

In this paper, we address the problem of achieving fast convergence in gradient descent for quadratic functions without relying on a priori knowledge of global function parameters. Inspired by adaptive stepsize algorithms for smooth convex functions, we propose a computationally lightweight strategy based on running estimates of minimal and maximal local curvatures. We prove that our proposed algorithm converges to the optimal constant stepsize which achieves the fastest convergence. Simulations show that the convergence rate achieved by our proposed algorithm is comparable or superior to recent adaptive approaches both in the quadratic case under consideration and in a preliminary test on logistic classification.

math.OC↗

arXiv2026

Adaptive Stepsizes With Certified Convergence in Distributed Gradient Tracking With Quadratic Costs

In this work, we propose an adaptive stepsize rule with guaranteed convergence for Distributed Gradient Tracking applied to scalar quadratic problems with heterogeneous curvatures. Most distributed gradient-based algorithms require a suitable stepsize selection. Available theoretical bounds are often overly conservative, while practical implementations typically rely on empirically tuned heuristics. Online adaptive strategies have only recently emerged for general distributed convex optimization, but their properties and performance remain only partially understood. To gain analytical insight, we focus on the informative setting of scalar quadratic costs, which allows us to explicitly capture the interplay between network topology and curvature heterogeneity. We derive a convergence bound parameterized only by the essential spectral radius of the consensus matrix and the heterogeneity of the local cost curvatures, both computable online without any prior knowledge of the optimization problem. Optimizing this bound yields a computationally tractable surrogate for the convergence rate and the optimal constant stepsize. The resulting stepsize admits an analytical interpretation, guarantees convergence for arbitrary network topologies and curvature heterogeneity, and is provably tight for complete graphs and homogeneous curvatures. Finally, extensive numerical simulations demonstrate that the proposed distributed adaptive strategy significantly outperforms existing offline and online stepsize selection rules in the considered setting.

math.OC↗

arXiv2026

Timescale Separation Through the Lens of Operator Theory

Timescale separation is a powerful tool for analyzing interconnected dynamical systems. Meanwhile, operator theory provides a general framework for studying the convergence of iterative methods formulated as fixed-point iterations, including algorithms arising in optimization, learning, and control. In this paper, we bridge these two areas by establishing timescale separation results for fixed-point iterations induced by both deterministic and stochastic operators. As customary in timescale separation, our results involve auxiliary systems that arise from the original interconnection in the limit as the timescale parameter tends to zero and separately capture the dynamics induced by the slow and fast operators. The proposed operator-theoretic framework yields explicit and readily checkable bounds on this tunable parameter, expressed in terms of standard operator constants. To illustrate the applicability of our results, we employ them to prove the convergence properties of a feedback optimization scheme in both deterministic and stochastic settings.

math.OC↗

arXiv2026

Submodular Multi-Agent Policy Learning for Online Distributed Task Allocation in Open Multi-Agent Systems

This paper studies multi-agent reinforcement learning with submodular team utilities for online distributed task allocation. In this setting, each agent selects one action from a local categorical policy, so feasible joint actions form a partition matroid over agent-action pairs. Classical multilinear extensions use independent Bernoulli sampling and therefore do not match the categorical policies executed by decentralized agents. To address this mismatch, we introduce the Partition Multilinear Extension (PME), a continuous relaxation whose value equals the expected team utility under factorized categorical policies. We prove that submodular difference rewards provide unbiased PME marginal-gradient information and yield a stagewise score-function policy-gradient estimator. Based on this connection, we propose SubMAPG, a centralized-training decentralized-execution policy-gradient framework with masked categorical policies and submodular difference-reward training signals. For the associated PME marginal-space projected stochastic-gradient dynamics, we prove a stagewise 1/2-approximation guarantee and sublinear dynamic regret in slowly varying environments, measured by the path length of the optimal PME marginals. To handle open systems with time-varying agents and targets, we instantiate SubMAPG with graph neural network policies. Experiments on multi-robot coverage and multi-target tracking show that SubMAPG outperforms local greedy and shared-reward baselines and is competitive with centralized myopic greedy strategies.

eess.SY↗

arXiv2026

Learning Structured Robot Policies from Vision-Language Models via Synthetic Neuro-Symbolic Supervision

Vision-Language Models (VLMs) have recently demonstrated strong capabilities in mapping multimodal observations to robot behaviors. However, most current approaches rely on end-to-end visuomotor policies that remain opaque and difficult to analyze, limiting their use in real-world robotic applications. In contrast, classical robotic systems often rely on structured policy representations that provide interpretability, modularity, and reactive execution. This work investigates how foundation models can be specialized to generate structured robot policies grounded in multimodal perception, bridging high-dimensional learning and symbolic control. We propose a neuro-symbolic approach in which a VLM synthesizes executable Behavior Tree policies from visual observations, natural language instructions, and structured system specifications. To enable scalable supervision without manual annotation, we introduce an automated pipeline that generates a synthetic multimodal dataset of domain-randomized scenes paired with instruction-policy examples produced by a foundation model. By decoupling structured task decomposition under constrained symbolic grammars from hardware-specific motor control, we demonstrate that a 12B-parameter model can learn structured spatial-symbolic mappings required for executable BT synthesis, solely through in-silico supervision. Real-world physical experiments on two heterogeneous robotic manipulators confirm that these structurally constrained policies achieve zero-shot transfer to real-world environments. The results emphasize that the data bottleneck in robotic planning can be bypassed by procedurally synthesizing high-fidelity, neuro-symbolic training data.

cs.RO↗

arXiv2026

Real2Sim via Active Perception with Behavior Trees Automatically Generated by VLMs

Constructing physically accurate simulation environments (Real2Sim) traditionally relies on manual system identification or rigid, exhaustive exploration routines. These task-agnostic pipelines often fail to leverage semantic scene context, leading to redundant physical interactions and inefficient data acquisition. In this paper, we present an autonomous, intent-driven Real2Sim framework that leverages Vision-Language Models (VLMs) for Semantic Task Decomposition. Given a high-level natural language request, an incomplete simulation description, and a visual observation, the framework autonomously identifies the minimal subset of missing physical parameters required for the simulation task. It then generates a reactive Behavior Tree (BT) composed of atomic motion and sensing primitives to selectively acquire these parameters through contact-rich robotic interaction. Extensive real-world experiments on a torque-controlled Franka Emika Panda demonstrate that our approach accurately estimates object mass, surface geometry, and derived parameters such as friction. Quantitative evaluations reveal significant operational efficiency gains compared to exhaustive baseline methods, while ablation studies confirm the robustness of the prompt architecture across different state-of-the-art VLMs. Furthermore, the reactive hierarchy of the BT acts as a deterministic safety filter, successfully mitigating generative VLM hallucinations and preventing unsafe physical anomalies. Ultimately, this work provides a scalable, efficient, and interpretable pipeline for building physics-aware digital twins directly from unstructured human intent.

cs.RO↗

arXiv2025

The Bayesian Separation Principle for Data-driven Control

In this paper we investigate the existence of a separation principle between model identification and control design in the context of model predictive control. First, we clarify that such a separation principle holds asymptotically in the number of data in a Fisherian context, and show that it holds universally, i.e. regardless of the data size, in a Bayesian context. Then, by formulating model predictive control within a Gaussian regression framework, we describe how the Bayesian separation principle can be used to derive computable, uncertainty-aware expressions for the control cost and optimal input sequence, thereby bridging direct and indirect data-driven approaches. Numerical results in both linear and nonlinear scenarios illustrate that the proposed approach outperform nominal methods that neglect uncertainty, highlighting the advantages of incorporating uncertainty in the control design process.

eess.SY↗

arXiv2025

Learning Stack-of-Tasks Management for Redundant Robots

This paper presents a novel framework for automatically learning complete Stack-of-Tasks (SoT) controllers for redundant robotic systems, including task priorities, activation logic, and control parameters. Unlike classical SoT pipelines-where task hierarchies are manually defined and tuned-our approach optimizes the full SoT structure directly from a user-specified cost function encoding intuitive preferences such as safety, precision, manipulability, or execution speed. The method combines Genetic Programming with simulation-based evaluation to explore both discrete (priority order, task activation) and continuous (gains, trajectory durations) components of the controller. We validate the framework on a dual-arm mobile manipulator (the ABB mobile-YuMi research platform), demonstrating robust convergence across multiple cost definitions, automatic suppression of irrelevant tasks, and strong resilience to distractors. Learned SoTs exhibit expert-like hierarchical structure and adapt naturally to multi-objective trade-offs. Crucially, all controllers transfer from Gazebo simulation to the real robot, achieving safe and precise motion without additional tuning. Experiments in static and dynamic environments show reliable obstacle avoidance, high tracking accuracy, and predictable behavior in the presence of humans. The proposed method provides an interpretable and scalable alternative to manual SoT design, enabling rapid, user-driven generation of task execution hierarchies for complex robotic systems.

cs.RO↗

arXiv2025

A Robust Controller based on Gaussian Processes for Robotic Manipulators with Unknown Uncertainty

In this paper, we propose a novel learning-based robust feedback linearization strategy to ensure precise trajectory tracking for an important family of Lagrangian systems. We assume a nominal knowledge of the dynamics is given but no a-priori bounds on the model mismatch are available. In our approach, the key ingredient is the adoption of a regression framework based on Gaussian Processes (GPR) to estimate the model mismatch. This estimate is added to the outer loop of a classical feedback linearization scheme based on the nominal knowledge available. Then, to compensate for the residual uncertainty, we robustify the controller including an additional term whose size is designed based on the variance provided by the GPR framework. We proved that, with high probability, the proposed scheme is able to guarantee asymptotic tracking of a desired trajectory. We tested numerically our strategy on a 2 degrees of freedom planar robot.

cs.RO↗

arXiv2025

Finetuning Deep Reinforcement Learning Policies with Evolutionary Strategies for Control of Underactuated Robots

Deep Reinforcement Learning (RL) has emerged as a powerful method for addressing complex control problems, particularly those involving underactuated robotic systems. However, in some cases, policies may require refinement to achieve optimal performance and robustness aligned with specific task objectives. In this paper, we propose an approach for fine-tuning Deep RL policies using Evolutionary Strategies (ES) to enhance control performance for underactuated robots. Our method involves initially training an RL agent with Soft-Actor Critic (SAC) using a surrogate reward function designed to approximate complex specific scoring metrics. We subsequently refine this learned policy through a zero-order optimization step employing the Separable Natural Evolution Strategy (SNES), directly targeting the original score. Experimental evaluations conducted in the context of the 2nd AI Olympics with RealAIGym at IROS 2024 demonstrate that our evolutionary fine-tuning significantly improves agent performance while maintaining high robustness. The resulting controllers outperform established baselines, achieving competitive scores for the competition tasks.

cs.RO↗

arXiv2025

Towards Autonomous Reinforcement Learning for Real-World Robotic Manipulation with Large Language Models

Recent advancements in Large Language Models (LLMs) and Visual Language Models (VLMs) have significantly impacted robotics, enabling high-level semantic motion planning applications. Reinforcement Learning (RL), a complementary paradigm, enables agents to autonomously optimize complex behaviors through interaction and reward signals. However, designing effective reward functions for RL remains challenging, especially in real-world tasks where sparse rewards are insufficient and dense rewards require elaborate design. In this work, we propose Autonomous Reinforcement learning for Complex Human-Informed Environments (ARCHIE), an unsupervised pipeline leveraging GPT-4, a pre-trained LLM, to generate reward functions directly from natural language task descriptions. The rewards are used to train RL agents in simulated environments, where we formalize the reward generation process to enhance feasibility. Additionally, GPT-4 automates the coding of task success criteria, creating a fully automated, one-shot procedure for translating human-readable text into deployable robot skills. Our approach is validated through extensive simulated experiments on single-arm and bi-manual manipulation tasks using an ABB YuMi collaborative robot, highlighting its practicality and effectiveness. Tasks are demonstrated on the real robot setup.

cs.RO↗

arXiv2025

Accelerating Model-Based Reinforcement Learning using Non-Linear Trajectory Optimization

This paper addresses the slow policy optimization convergence of Monte Carlo Probabilistic Inference for Learning Control (MC-PILCO), a state-of-the-art model-based reinforcement learning (MBRL) algorithm, by integrating it with iterative Linear Quadratic Regulator (iLQR), a fast trajectory optimization method suitable for nonlinear systems. The proposed method, Exploration-Boosted MC-PILCO (EB-MC-PILCO), leverages iLQR to generate informative, exploratory trajectories and initialize the policy, significantly reducing the number of required optimization steps. Experiments on the cart-pole task demonstrate that EB-MC-PILCO accelerates convergence compared to standard MC-PILCO, achieving up to $\bm{45.9\%}$ reduction in execution time when both methods solve the task in four trials. EB-MC-PILCO also maintains a $\bm{100\%}$ success rate across trials while solving the task faster, even in cases where MC-PILCO converges in fewer iterations.

cs.LG↗

arXiv2025

Learning global control of underactuated systems with Model-Based Reinforcement Learning

This short paper describes our proposed solution for the third edition of the "AI Olympics with RealAIGym" competition, held at ICRA 2025. We employed Monte-Carlo Probabilistic Inference for Learning Control (MC-PILCO), an MBRL algorithm recognized for its exceptional data efficiency across various low-dimensional robotic tasks, including cart-pole, ball \& plate, and Furuta pendulum systems. MC-PILCO optimizes a system dynamics model using interaction data, enabling policy refinement through simulation rather than direct system data optimization. This approach has proven highly effective in physical systems, offering greater data efficiency than Model-Free (MF) alternatives. Notably, MC-PILCO has previously won the first two editions of this competition, demonstrating its robustness in both simulated and real-world environments. Besides briefly reviewing the algorithm, we discuss the most critical aspects of the MC-PILCO implementation in the tasks at hand: learning a global policy for the pendubot and acrobot systems.

cs.RO↗

arXiv2025

Data-driven Power Loss Identification through Physics-Based Thermal Model Backpropagation

Digital twins for power electronics require accurate power losses whose direct measurements are often impractical or impossible in real-world applications. This paper presents a novel hybrid framework that combines physics-based thermal modeling with data-driven techniques to identify and correct power losses accurately using only temperature measurements. Our approach leverages a cascaded architecture where a neural network learns to correct the outputs of a nominal power loss model by backpropagating through a reduced-order thermal model. We explore two neural architectures, a bootstrapped feedforward network, and a recurrent neural network, demonstrating that the bootstrapped feedforward approach achieves superior performance while maintaining computational efficiency for real-time applications. Between the interconnection, we included normalization strategies and physics-guided training loss functions to preserve stability and ensure physical consistency. Experimental results show that our hybrid model reduces both temperature estimation errors (from 7.2+-6.8{\deg}C to 0.3+-0.3{\deg}C) and power loss prediction errors (from 5.4+-6.6W to 0.2+-0.3W) compared to traditional physics-based approaches, even in the presence of thermal model uncertainties. This methodology allows us to accurately estimate power losses without direct measurements, making it particularly helpful for real-time industrial applications where sensor placement is hindered by cost and physical limitations.

eess.SY↗