arXiv subjects

Eduardo Sebastián

Publications and source records attributed to Eduardo Sebastián.

15 recordsLinked to original sources

arXiv2026

Prompting Robot Teams with Natural Language

This paper presents a framework to prompt multi-robot teams with high-level tasks using natural language expressions. Our objective is to use the reasoning capabilities of language models in understanding and decomposing multi-robot collaboration and decision-making tasks, but in settings where such models cannot be called at deployment time. However, it is hard to specify the behavior of an individual robot from a team instruction, and have it continuously adapt to actions from other robots. This necessitates a framework with the representational capacity required by the logic and semantics of a task, and yet supports decentralized, real-time operation. We solve this dilemma by recognizing that a task can be represented as a deterministic finite automaton, and that recurrent neural networks (RNNs) can encode numerous automata. This allows us to distill the logic and sequential decompositions of sub-tasks obtained from a language model into an RNN, and align its internal states with the semantics of a given task. This leads to a tiny model that encapsulates the reasoning of the language model and can be implemented onboard. To interpret the internal state of the RNN for a decentralized execution, we train a graph neural network control policy conditioned on the hidden states of the RNN and the language embeddings. We present evaluations on simulated and real-world multi-robot tasks that require sequential and collaborative behavior by the team, demonstrating scalable, robust, real-time performance -- sites.google.com/view/prompting-teams.

cs.RO↗

arXiv2026

Policy Gradient with Self-Attention for Model-Free Distributed Nonlinear Multi-Agent Games

Multi-agent games in dynamic nonlinear settings are challenging due to the time-varying interactions among the agents and the non-stationarity of the (potential) Nash equilibria. In this paper we consider model-free games, where agent transitions and costs are observed without knowledge of the transition and cost functions that generate them. We propose a novel distributed policy structure that follows the communication constraints in multi-team games, with multiple agents per team, and learned through policy gradients. Our formulation is inspired by the structure of distributed policies in linear quadratic games, which take the form of time-varying linear feedback gains. In the nonlinear case, we model the policies as nonlinear feedback gains, parameterized by self-attention layers to account for the time-varying multi-agent communication topology. We demonstrate that our approach achieves strong performance in several settings, including distributed linear and nonlinear regulation, and simulated and real multi-robot pursuit-and-evasion games.

eess.SY↗

arXiv2026

World-Task Factorization for Robot Learning

Robot learning must produce policies that generalize to new combinations of constraints, teammates, and environments. To achieve this, we must structurally factor the policy, which is a choice that dictates what generalizes, what requires retraining, and what remains entangled. Existing methods span a wide spectrum, from expecting structure to emerge from data scaling, to hand-designing it via hierarchies, skill libraries or learned specializations. In this paper, we study what we argue is the most fundamental factorization in robotics: separating the world from the task. We investigate the conditions under which this factorization is principled. World factors are properties of the embodied system and the environment; they exist independently of intent. Task factors are defined by the task's logic over what the world admits. We formalize this asymmetry through Bayesian model evidence: it aligns with the data-generating process, maintains high likelihood through an analytical world model, and reduces the Occam razor's penalty on task parameters. We instantiate this factorization by pairing AICON, a differentiable graph of recursive estimators and interconnections that is compositional, operates without task-specific data, and propagates cost gradients to actuators, with a compact, learned policy that modulates gradient paths. Gradients serve as the interface between the two factors: they carry world structure through the graph and task structure through costs, enabling low-dimensional learning while preserving structural generalization. We test the world/task factorization across three problems that encompass heterogeneous robots, environments, task logic and sensorimotor modalities. Our framework outperforms end-to-end baselines and analytical heuristics in all settings, generalizes zero-shot to out-of-distribution configurations, and transfers to real hardware without retraining.

cs.RO↗

arXiv2026

Events as Triggers for Behavioral Diversity in Multi-Agent Reinforcement Learning

Effective multi-agent cooperation requires agents to adopt diverse behaviors as task conditions evolve-and to do so at the right moment. Yet, current Multi-Agent Reinforcement Learning (MARL) frameworks that facilitate this diversity are still limited by the fact that they bind fixed behaviors to fixed agent identities. Consequently, they are ill-equipped for tasks where agents need to take on different roles at very specific moments in time. We argue that, to define these behavioral transitions, the missing ingredient is $\textbf{events}$. Events are changes in the state of the system that induce qualitative changes in the task. Based on this view, we introduce a framework that decouples agent identity from behavior, capturing a continuous manifold from which agents instantiate their behaviors in response to events. This framework is based on two elements. First, to build an expressive behavior manifold, we introduce Neural Manifold Diversity (NMD), a formal distance metric that remains well-defined when behaviors are transient and agent-agnostic. Second, we use an event-based hypernetwork that generates Low-Rank Adaptation (LoRA) modules over a shared team policy, enabling on-the-fly agent-policy reconfiguration in response to events. We prove that this construction ensures that diversity does not interfere with reward maximization by design. Empirical results demonstrate that our framework outperforms established baselines across benchmarks while exhibiting zero-shot generalization, and being the only method that solves tasks requiring sequential behavior reassignment.

cs.MA↗

arXiv2026

Wake Up to the Past: Using Memory to Model Fluid Wake Effects on Robots

Autonomous aerial and aquatic robots that attain mobility by perturbing their medium, such as multicopters and torpedoes, produce wake effects that act as disturbances for adjacent robots. Wake effects are hard to model and predict due to the chaotic spatio-temporal dynamics of the fluid, entangled with the physical geometry of the robots and their complex motion patterns. Data-driven approaches using neural networks typically learn a memory-less function that maps the current states of the two robots to a force observed by the "sufferer" robot. Such models often perform poorly in agile scenarios: since the wake effect has a finite propagation time, the disturbance observed by a sufferer robot is some function of relative states in the past. In this work, we present an empirical study of the properties a wake-effect predictor must satisfy to accurately model the interactions between two robots mediated by a fluid. We explore seven data-driven models designed to capture the spatio-temporal evolution of fluid wake effects in four different media. This allows us to introspect the models and analyze the reasons why certain features enable improved accuracy in prediction across predictors and fluids. As experimental validation, we develop a planar rectilinear gantry for two spinning monocopters to test in real-world data with feedback control. The conclusion is that support of history of previous states as input and transport delay prediction substantially helps to learn an accurate wake-effect predictor.

cs.RO↗

arXiv2026

From Vision to Decision: Neuromorphic Control for Autonomous Navigation and Tracking

Robotic navigation has historically struggled to reconcile reactive, sensor-based control with the decisive capabilities of model-based planners. This duality becomes critical when the absence of a predominant option among goals leads to indecision, challenging reactive systems to break symmetries without computationally-intense planners. We propose a parsimonious neuromorphic control framework that bridges this gap for vision-guided navigation and tracking. Image pixels from an onboard camera are encoded as inputs to dynamic neuronal populations that directly transform visual target excitation into egocentric motion commands. A dynamic bifurcation mechanism resolves indecision by delaying commitment until a critical point induced by the environmental geometry. Inspired by recently proposed mechanistic models of animal cognition and opinion dynamics, the neuromorphic controller provides real-time autonomy with a minimal computational burden, a small number of interpretable parameters, and can be seamlessly integrated with application-specific image processing pipelines. We validate our approach in simulation environments as well as on an experimental quadrotor platform.

cs.RO↗

arXiv2025

Curriculum Imitation Learning of Distributed Multi-Robot Policies

Learning control policies for multi-robot systems (MRS) remains a major challenge due to long-term coordination and the difficulty of obtaining realistic training data. In this work, we address both limitations within an imitation learning framework. First, we shift the typical role of Curriculum Learning in MRS, from scalability with the number of robots, to focus on improving long-term coordination. We propose a curriculum strategy that gradually increases the length of expert trajectories during training, stabilizing learning and enhancing the accuracy of long-term behaviors. Second, we introduce a method to approximate the egocentric perception of each robot using only third-person global state demonstrations. Our approach transforms idealized trajectories into locally available observations by filtering neighbors, converting reference frames, and simulating onboard sensor variability. Both contributions are integrated into a physics-informed technique to produce scalable, distributed policies from observations. We conduct experiments across two tasks with varying team sizes and noise levels. Results show that our curriculum improves long-term accuracy, while our perceptual estimation method yields policies that are robust to realistic uncertainty. Together, these strategies enable the learning of robust, distributed controllers from global demonstrations, even in the absence of expert actions or onboard measurements.

cs.RO↗

arXiv2025

LATMOS: Latent Automaton Task Model from Observation Sequences

Robot task planning from high-level instructions is an important step towards deploying fully autonomous robot systems in the service sector. Three key aspects of robot task planning present challenges yet to be resolved simultaneously, namely, (i) factorization of complex tasks specifications into simpler executable subtasks, (ii) understanding of the current task state from raw observations, and (iii) planning and verification of task executions. To address these challenges, we propose LATMOS, an automata-inspired task model that, given observations from correct task executions, is able to factorize the task, while supporting verification and planning operations. LATMOS combines an observation encoder to extract the features from potentially high-dimensional observations with automata theory to learn a sequential model that encapsulates an automaton with symbols in the latent feature space. We conduct extensive evaluations in three task model learning setups: (i) abstract tasks described by logical formulas, (ii) real-world human tasks described by videos and natural language prompts and (iii) a robot task described by image and state observations. The results demonstrate the improved plan generation and verification capabilities of LATMOS across observation modalities and tasks.

cs.RO↗

arXiv2025

AVOCADO: Adaptive Optimal Collision Avoidance driven by Opinion

We present AVOCADO (AdaptiVe Optimal Collision Avoidance Driven by Opinion), a novel navigation approach to address holonomic robot collision avoidance when the robot does not know how cooperative the other agents in the environment are. AVOCADO departs from a Velocity Obstacle's (VO) formulation akin to the Optimal Reciprocal Collision Avoidance method. However, instead of assuming reciprocity, it poses an adaptive control problem to adapt to the cooperation level of other robots and agents in real time. This is achieved through a novel nonlinear opinion dynamics design that relies solely on sensor observations. As a by-product, we leverage tools from the opinion dynamics formulation to naturally avoid the deadlocks in geometrically symmetric scenarios that typically suffer VO-based planners. Extensive numerical simulations show that AVOCADO surpasses existing motion planners in mixed cooperative/non-cooperative navigation environments in terms of success rate, time to goal and computational time. In addition, we conduct multiple real experiments that verify that AVOCADO is able to avoid collisions in environments crowded with other robots and humans.

cs.RO↗

arXiv2025

Distributed Discrete-time Dynamic Outer Approximation of the Intersection of Ellipsoids

This paper presents the first discrete-time distributed algorithm to track the tightest ellipsoids that outer approximates the global dynamic intersection of ellipsoids. Given an undirected network, we consider a setup where each node measures an ellipsoid, defined as a time-varying positive semidefinite matrix. The goal is to devise a distributed algorithm to track the tightest outer approximation of the intersection of all the ellipsoids. The solution is based on a novel distributed reformulation of the original centralized semi-definite outer Löwner-John program, characterized by a non-separable objective function and global constraints. We prove finite-time convergence to the global minima of the centralized problem in the static case and finite-time bounded tracking error in the dynamic case. Moreover, we prove boundedness of estimation in the tracking of the global optimum and robustness in the estimation against time-varying inputs. We illustrate the properties of the algorithm with different simulated examples, including a distributed estimation showcase where our proposal is integrated into a distributed Kalman filter to surpass the state-of-the-art in mean square error performance.

math.OC↗

arXiv2024

Accelerated Alternating Direction Method of Multipliers Gradient Tracking for Distributed Optimization

This paper presents a novel accelerated distributed algorithm for unconstrained consensus optimization over static undirected networks. The proposed algorithm combines the benefits of acceleration from momentum, the robustness of the alternating direction method of multipliers, and the computational efficiency of gradient tracking to surpass existing state-of-the-art methods in convergence speed, while preserving their computational and communication cost. First, we prove that, by applying momentum on the average dynamic consensus protocol over the estimates and gradient, we can study the algorithm as an interconnection of two singularly perturbed systems: the outer system connects the consensus variables and the optimization variables, and the inner system connects the estimates of the optimum and the auxiliary optimization variables. Next, we prove that, by adding momentum to the auxiliary dynamics, our algorithm always achieves faster convergence than the achievable linear convergence rate for the non-accelerated alternating direction method of multipliers gradient tracking algorithm case. Through simulations, we numerically show that our accelerated algorithm surpasses the existing accelerated and non-accelerated distributed consensus first-order optimization protocols in convergence speed.

math.OC↗

arXiv2023

ECO-DKF: Event-Triggered and Certifiable Optimal Distributed Kalman Filter under Unknown Correlations

This paper presents ECO-DKF, the first Event-Triggered and Certifiable Optimal Distributed Kalman Filter. Our algorithm addresses two major issues inherent to Distributed Kalman Filters: (i) fully distributed and scalable optimal estimation and (ii) reduction of the communication bandwidth usage. The first requires to solve an NP-hard optimisation problem, forcing relaxations that lose optimality guarantees over the original problem. Using only information from one-hop neighbours, we propose a tight Semi-Definite Programming relaxation that allows to certify locally and online if the relaxed solution is the optimum of the original NP-hard problem. In that case, ECO-DKF is optimal in the square error sense under scalability and event-triggered one-hop communications restrictions. Additionally, ECO-DKF is a globally asymptotically stable estimator. To address the second issue, we propose an event-triggered scheme from the relaxed optimisation output. The consequence is a broadcasting-based algorithm that saves communication bandwidth, avoids individual communication links and multiple information exchanges within instants, and preserves the optimality and stability properties of the filter.

eess.SY↗

arXiv2022

Multi-robot Implicit Control of Herds

This paper presents a novel control strategy to herd a group of non-cooperative evaders by means of a team of robotic herders. In herding problems, the motion of the evaders is typically determined by strong nonlinear reactive dynamics, escaping from the herders. Many applications demand the herding of numerous and/or heterogeneous entities, making the development of flexible control solutions challenging. In this context, our main contribution is a control approach that finds suitable herding actions even when the nonlinearities in the evaders' dynamics yield to implicit equations. We resort to numerical analysis theory to characterise the existence conditions of such actions and propose two design methods to compute them, one transforming the continuous time implicit system into an expanded explicit system, and the other applying a numerical method to find the action in discrete time. Simulations and real experiments validate the proposal in different scenarios.

eess.SY↗

arXiv2022

All-in-one: Certifiable Optimal Distributed Kalman Filter under Unknown Correlations

The optimal fusion of estimates in a Distributed Kalman Filter (DKF) requires tracking of the complete network error covariance, problematic in terms of memory and communication. A scalable alternative is to fuse estimates under unknown correlations, doing the update by solving an optimisation problem. Unfortunately, this problem is NP-hard, forcing relaxations that lose optimality guarantees. Motivated by this, we present the first Certifiable Optimal DKF (CO-DKF). Using only information from one-hop neighbours, CO-DKF solves the optimal fusion of estimates under unknown correlations by a particular tight Semidefinite Programming (SDP) relaxation which allows to certify, locally and in real time, if the relaxed solution is the actual optimum. In that case, we prove optimality in the Mean Square Error (MSE) sense. Additionally, we demonstrate the global asymptotic stability of the estimator. CO-DKF outperforms other state-of-the-art DKF algorithms, specially in sparse, highly noisy setups.

eess.SY↗

arXiv2022

Adaptive Multi-robot Implicit Control of Heterogeneous Herds

This paper presents a novel control strategy to herd groups of non-cooperative evaders by means of a team of robotic herders. In herding problems, the motion of the evaders is typically determined by strongly nonlinear and heterogeneous reactive dynamics, which makes the development of flexible control solutions a challenging problem. In this context, we propose Implicit Control, an approach that leverages numerical analysis theory to find suitable herding inputs even when the nonlinearities in the evaders' dynamics yield to implicit equations. The intuition behind this methodology consists in driving the input, rather than computing it, towards the unknown value that achieves the desired dynamic behavior of the herd. The same idea is exploited to develop an adaptation law, with stability guarantees, that copes with uncertainties in the herd's models. Moreover, our solution is completed with a novel caging technique based on uncertainty models and Control Barrier Functions (CBFs), together with a distributed estimator to overcome the need of complete perfect measurements. Different simulations and experiments validate the generality and flexibility of the proposal.

eess.SY↗