arXiv subjects

Victor Barasuol

Publications and source records attributed to Victor Barasuol.

At least 19 recordsLinked to original sources

arXiv2026

BeyondContact: A Reinforcement Learning Framework for Airflow-Based Non-Contact Manipulation

Contact-based robotic manipulation is constrained by end-effector reachability and the mechanical properties of objects. Airflow can exert forces without direct robot-object contact, enabling manipulation of objects beyond the reach or when direct contact is undesirable. However, the lack of suitable simulation has hindered the application of reinforcement learning to airflow-based manipulation, limiting the broader adoption of this mechanism compared with contact-based methods. To address this gap, we developed a robot- and task-agnostic airflow extension for Isaac Lab that supports massively parallel training. The extension uses experimentally identified airflow models to compute the forces and torques exerted on objects. Based on this extension, we train a hybrid push-blow (HPB) policy that combines contact-based pushing with airflow manipulation and a rectangular reorientation (RR) policy that controls a cuboid's heading using airflow alone. To enable physical execution, we design a lightweight, high-speed blower that coexists with a gripper. In zero-shot real-world deployment, HPB achieves an 80% success rate over 20 trials, while RR achieves a mean absolute heading error of 2.6 degrees across four target headings. The project materials are available at https://github.com/iit-DLSLab/BeyondContact.

cs.RO↗

arXiv2026

Learning from Relative Action Credit for Escaping Local Minima

Reinforcement learning holds the promise of robots that learn manipulation skills from interaction alone, without pretrained priors or demonstration data. In practice, on-policy methods such as PPO routinely stall on long-horizon, contact-rich tasks: dense reward shaping carries the policy through stages, yet stitching them together and discovering sparsely rewarded events rarely succeeds. We trace this failure to the missing action discrimination in state-value credit assignment: a state-value critic judges progress only by episode outcomes, so when the policy converges to a partial solution all outcomes coincide, and improvement stops because no action can prove better than another. We therefore extend PPO with an action conditioned value critic which, can compare the executed action with alternatives sampled at the same state and aid the learning signal. Combined with Generalized Advantage Estimation and a bound on exploration noise, we show how the extended algorithm escapes the plateau and completes precise peg insertion within a training budget where PPO with identical exploration fails.

cs.LG↗

arXiv2026

ZiMPedance: Impedance-Aware ZMP Modeling and Control for Payload Carrying with Quadruped Robots

Load transportation with quadruped robots is strongly affected by the dynamics of the physical interface between the robot and the load. Passive spring-based arms provide a lightweight and mechanically compliant alternative to active manipulators, but their dynamics can store and release energy during locomotion, producing oscillatory interaction forces that reduce robot stability. This paper develops a passive-arm-aware ZMP model for a quadruped carrying a payload through a passive arm. The reduced analytical model relates passive-interface stiffness, damping, and payload mass to the ZMP response under both external payload forcing and locomotion-induced base excitation, showing that lightly damped passive modes can overlap with gait harmonics. We then propose an arm-aware MPC (ARMPC), a passive-arm Model Predictive Control approach based on extended Single Rigid Body Dynamics that combines reduced floating-base dynamics with passive-arm joint states, internal torsional springs and dampers, and base-arm inertial coupling. Simulation ablations show that ARMPC consistently reduces ZMP-margin violations and prevents failures across all tested conditions, while Nominal MPC, which neither models nor compensates the passive arm, and Baseline MPC, which compensates it through an estimated wrench, exhibit higher violation percentages and failures under disturbances. Hardware pull-release experiments with a 2 kg payload show that the proposed ARMPC completes all longitudinal and diagonal disturbance trials without failure, whereas Baseline MPC fails in 7/15 and 8/15 trials, respectively. The same model also enables passive-arm end-effector tracking while maintaining a positive ZMP margin.

cs.RO↗

arXiv2026

From Tabletop to Walking Bot: Deploying a VLA Model on a Legged Manipulator Robot

Vision-language-action (VLA) models are typically pretrained on broad, diverse datasets and subsequently fine-tuned on a comparatively small dataset tailored to a specific target robot. Such adaptation recipe is now widely established, achieving its most mature results on fixed-base arms due to their static base dynamics and consistent reference frame. Mounting the same arm on a quadruped extends the workspace to the entire floor, but fundamentally changes the embodiment: rather than controlling the arm in isolation, the policy must jointly coordinate base navigation, body posture, and end-effector motion. In this work, we study the practical demands of this embodiment change across data collection, model adaptation, and inference design to achieve closed-loop loco-manipulation on a Unitree Go2 with an AgileX Piper arm. We report the choices that made this real-robot deployment feasible, namely model variant, action horizon, and demonstration conventions, together with a study tailored to a pick-and-place task, separating data coverage from model capacity. We show that expanding demonstration coverage reduces held-out end-effector error by 30%, and incorporating secondary skills (such as pushing rather than grasping) introduces new behavioral modes without degrading baseline performance. Finally, closed-loop real-world experiments show how dataset composition shapes success and failure, validating the pipeline for more complex loco-manipulation.

cs.RO↗

arXiv2026

LocoPilot: Single-Operator Bilateral Teleoperation for Loco-Manipulation Policy Learning

Learning loco-manipulation policies requires demonstrations that capture coordinated base and arm motion on the target embodiment. Robot-free interfaces allow demonstrations to be collected independently of the target robot, increasing flexibility but without directly capturing platform-specific commands, proprioception, and body-mounted observations. On-robot teleoperation allows to capture these signals directly, but requires an interface capable of commanding the complete system concurrently. This paper presents LocoPilot, a single-operator bilateral teleoperation and data-collection framework implemented on a 12-DoF Unitree Go2 quadruped equipped with a 6-DoF AgileX Piper L manipulator, an Intel RealSense D435 wrist camera, and an Intel RealSense D405 head camera. A second 6-DoF Piper L arm, used as the leader and two handheld controllers provide simultaneous control of the base, body posture, manipulator, and gripper. The follower arm interaction forces are estimated without a dedicated force/torque sensor and rendered to the operator through a passivity-constrained haptic channel. A ROS~2 pipeline records synchronized multimodal observations together with a unified 11D action shared by teleoperation and policy deployment. Compared to prior work, LocoPilot uniquely unifies robot-in-the-loop legged data collection, force feedback without motion-capture instrumentation, and the same action interface at deployment. We evaluate the interface across multiple loco-manipulation tasks and validate the complete collection-to-deployment pipeline by fine-tuning MolmoAct2 \cite{MOLMO} on pick-and-place demonstrations and deploying on the physical robot without action retargeting or additional inverse kinematics. Code available at: https://github.com/iit-DLSLab/LocoPilot

cs.RO↗

arXiv2026

Terrain-Aware Autonomous Planetary Exploration for Exteroceptive-Proprioceptive Mapping with Quadruped Scouts

Autonomous planetary exploration requires robots to navigate unknown, uneven terrain while assessing risk, traversability, and energetic cost. Quadruped scouts are well suited for this task because they can traverse irregular surfaces and gather mobility-relevant information during locomotion. This paper presents a terrain-aware exploration framework that combines exteroceptive and proprioceptive mapping for a quadruped robot in lunar-like environments. An onboard RGB-D camera builds robot-centered elevation maps, estimates geometric traversability, and derives navigation costs for autonomous planning. In parallel, proprioceptive measurements provide interaction-aware terrain cues that complement geometry-based assessment. Local maps are incrementally registered into a global multi-layer representation, which is used by an exploration module to select targets in unexplored regions of interest. The targets are reached by an autonomous navigation system that guides collision-aware motion using the available map and cost layers. Simulation results on NVIDIA Isaac Sim show autonomous exploration, map expansion, and spatial association between terrain geometry and robot-terrain interaction. Subsequent navigation using this information exhibits lower average Cost of Transport (CoT) than initial exploration.

cs.RO↗

arXiv2026

Gait-Dependent Effects on Quadruped Locomotion for Load-Carrying using Passive Mechanism

Passive mechanical interfaces offer a lightweight alternative to actuated manipulators for quadruped payload carrying, but their impedance directly couples the payload dynamics with the locomotion pattern. This paper analyzes how passive-arm stiffness-damping selection affects payload-carrying locomotion under different gait and payload conditions. We compare damped and underdamped passive-arm impedance configurations in simulation during flat-ground locomotion. For crawl gaits, where the support polygon remains well defined, the results show that underdamped impedance increases passive-joint oscillations and can reduce the ZMP margin with respect to the support polygon. Trot is retained as a dynamic excitation case for the passive arm, but it is not used for direct ZMP-margin stability comparison. The results are summarized in gait-payload-stiffness-damping maps, where ZMP-margin reduction is evaluated for crawl gaits and trot is retained only as a passive-arm excitation case.

cs.RO↗

arXiv2026

Proprioceptive Image: An Image Representation of Proprioceptive Data from Quadruped Robots for Contact Estimation Learning

This paper presents a novel approach for representing proprioceptive time-series data from quadruped robots as structured two-dimensional images, enabling the use of convolutional neural networks for learning locomotion-related tasks. The proposed method encodes temporal dynamics from multiple proprioceptive signals, such as joint positions, IMU readings, and foot velocities, while preserving the robot's morphological structure in the spatial arrangement of the image. This transformation captures inter-signal correlations and gait-dependent patterns, providing a richer feature space than direct time-series processing. We apply this concept in the problem of contact estimation, a key capability for stable and adaptive locomotion on diverse terrains. Experimental evaluations on both real-world datasets and simulated environments show that our image-based representation consistently enhances prediction accuracy and generalization over conventional sequence-based models, underscoring the potential of cross-modal encoding strategies for robotic state learning. Our method achieves superior performance on the contact dataset, improving contact state accuracy from 87.7% to 94.5% over the recently proposed MI-HGNN method, using a 15 times shorter window size.

cs.RO↗

arXiv2026

CroSTAta: Cross-State Transition Attention Transformer for Robotic Manipulation

Learning robotic manipulation policies through supervised learning from demonstrations remains challenging when policies encounter execution variations not explicitly covered during training. While incorporating historical context through attention mechanisms can improve robustness, standard approaches process all past states in a sequence without explicitly modeling the temporal structure that demonstrations may include, such as failure and recovery patterns. We propose a Cross-State Transition Attention Transformer that employs a novel State Transition Attention (STA) mechanism to modulate standard attention weights based on learned state evolution patterns, enabling policies to better adapt their behavior based on execution history. Our approach combines this structured attention with temporal masking during training, where visual information is randomly removed from recent timesteps to encourage temporal reasoning from historical context. Evaluation in simulation shows that STA consistently outperforms standard attention approach and temporal modeling methods like TCN and LSTM networks, achieving more than 2x improvement over cross-attention on precision-critical tasks. The source code and data can be accessed at https://github.com/iit-DLSLab/croSTAta

cs.RO↗

arXiv2026

Floating-Base Deep Lagrangian Networks

Grey-box methods for system identification combine deep learning with physics-informed constraints, capturing complex dependencies while improving out-of-distribution generalization. Despite the growing importance of floating-base systems such as humanoids and quadrupeds, current grey-box models ignore their specific physical constraints. For instance, the inertia matrix is not only positive definite but also exhibits branch-induced sparsity and input independence. Moreover, the 6x6 composite spatial inertia of the floating base inherits properties of single-rigid-body inertia matrices. As we show, this includes the triangle inequality on the eigenvalues of the composite rotational inertia. To address the lack of physical consistency in deep learning models of floating-base systems, we introduce a parameterization of inertia matrices that satisfies all these constraints. Inspired by Deep Lagrangian Networks (DeLaN), we train neural networks to predict physically plausible inertia matrices that minimize inverse dynamics error under Lagrangian mechanics. For evaluation, we collected and released a dataset on multiple quadrupeds and humanoids. In these experiments, our Floating-Base Deep Lagrangian Networks (FeLaN) achieve better overall performance on both simulated and real robots, while providing greater physical interpretability.

cs.RO↗

arXiv2025

Load-Based Variable Transmission Mechanism for Robotic Applications

This paper presents a Load-Based Variable Transmission (LBVT) mechanism designed to enhance robotic actuation by dynamically adjusting the transmission ratio in response to external torque demands. Unlike existing variable transmission systems that require additional actuators for active control, the proposed LBVT mechanism leverages a pre-tensioned spring and a four-bar linkage to passively modify the transmission ratio, thereby reducing the complexity of robot joint actuation systems. The effectiveness of the LBVT mechanism is evaluated through simulation-based analyses. The results confirm that the system achieves up to a 40 percent increase in transmission ratio upon reaching a predefined torque threshold, effectively amplifying joint torque when required without additional actuation. Furthermore, the simulations demonstrate a torque amplification effect triggered when the applied force exceeds 18 N, highlighting the system ability to autonomously respond to varying load conditions. This research contributes to the development of lightweight, efficient, and adaptive transmission systems for robotic applications, particularly in legged robots where dynamic torque adaptation is critical.

cs.RO↗

arXiv2025

Primal-Dual iLQR for GPU-Accelerated Learning and Control in Legged Robots

This paper introduces a novel Model Predictive Control (MPC) implementation for legged robot locomotion that leverages GPU parallelization. Our approach enables both temporal and state-space parallelization by incorporating a parallel associative scan to solve the primal-dual Karush-Kuhn-Tucker (KKT) system. In this way, the optimal control problem is solved in $\mathcal{O}(\log^2(n)\log{N} + \log^2(m))$ complexity, instead of $\mathcal{O}(N(n + m)^3)$, where $n$, $m$, and $N$ are the dimension of the system state, control vector, and the length of the prediction horizon. We demonstrate the advantages of this implementation over two state-of-the-art solvers (acados and crocoddyl), achieving up to a 60\% improvement in runtime for Whole Body Dynamics (WB)-MPC and a 700\% improvement for Single Rigid Body Dynamics (SRBD)-MPC when varying the prediction horizon length. The presented formulation scales efficiently with the problem state dimensions as well, enabling the definition of a centralized controller for up to 16 legged robots that can be computed in less than 25 ms. Furthermore, thanks to the JAX implementation, the solver supports large-scale parallelization across multiple environments, allowing the possibility of performing learning with the MPC in the loop directly in GPU. The code associated with this work can be found at https://github.com/iit-DLSLab/mpx.

cs.RO↗

arXiv2025

Non-Gaited Legged Locomotion with Monte-Carlo Tree Search and Supervised Learning

Legged robots are able to navigate complex terrains by continuously interacting with the environment through careful selection of contact sequences and timings. However, the combinatorial nature behind contact planning hinders the applicability of such optimization problems on hardware. In this work, we present a novel approach that optimizes gait sequences and respective timings for legged robots in the context of optimization-based controllers through the use of sampling-based methods and supervised learning techniques. We propose to bootstrap the search by learning an optimal value function in order to speed-up the gait planning procedure making it applicable in real-time. To validate our proposed method, we showcase its performance both in simulation and on hardware using a 22 kg electric quadruped robot. The method is assessed on different terrains, under external perturbations, and in comparison to a standard control approach where the gait sequence is fixed a priori.

cs.RO↗

arXiv2025

Load-bearing Assessment for Safe Locomotion of Quadruped Robots on Collapsing Terrain

Collapsing terrains, often present in search and rescue missions or planetary exploration, pose significant challenges for quadruped robots. This paper introduces a robust locomotion framework for safe navigation over unstable surfaces by integrating terrain probing, load-bearing analysis, motion planning, and control strategies. Unlike traditional methods that rely on specialized sensors or external terrain mapping alone, our approach leverages joint measurements to assess terrain stability without hardware modifications. A Model Predictive Control (MPC) system optimizes robot motion, balancing stability and probing constraints, while a state machine coordinates terrain probing actions, enabling the robot to detect collapsible regions and dynamically adjust its footholds. Experimental results on custom-made collapsing platforms and rocky terrains demonstrate the framework's ability to traverse collapsing terrain while maintaining stability and prioritizing safety.

cs.RO↗

arXiv2025

Towards An Adaptive Locomotion Strategy For Quadruped Rovers: Quantifying When To Slide Or Walk On Planetary Slopes

Legged rovers provide enhanced mobility compared to wheeled platforms, enabling navigation on steep and irregular planetary terrains. However, traditional legged locomotion might be energetically inefficient and potentially dangerous to the rover on loose and inclined surfaces, such as crater walls and cave slopes. This paper introduces a preliminary study that compares the Cost of Transport (CoT) of walking and torso-based sliding locomotion for quadruped robots across different slopes, friction conditions and speed levels. By identifying intersections between walking and sliding CoT curves, we aim to define threshold conditions that may trigger transitions between the two strategies. The methodology combines physics-based simulations in Isaac Sim with particle interaction validation in ANSYS-Rocky. Our results represent an initial step towards adaptive locomotion strategies for planetary legged rovers.

cs.RO↗

arXiv2025

Towards Proprioceptive Terrain Mapping with Quadruped Robots for Exploration in Planetary Permanently Shadowed Regions

Permanently Shadowed Regions (PSRs) near the lunar poles are of interest for future exploration due to their potential to contain water ice and preserve geological records. Their complex, uneven terrain favors the use of legged robots, which can traverse challenging surfaces while collecting in-situ data, and have proven effective in Earth analogs, including dark caves, when equipped with onboard lighting. While exteroceptive sensors like cameras and lidars can capture terrain geometry and even semantic information, they cannot quantify its physical interaction with the robot, a capability provided by proprioceptive sensing. We propose a terrain mapping framework for quadruped robots, which estimates elevation, foot slippage, energy cost, and stability margins from internal sensing during locomotion. These metrics are incrementally integrated into a multi-layer 2.5D gridmap that reflects terrain interaction from the robot's perspective. The system is evaluated in a simulator that mimics a lunar environment, using the 21 kg quadruped robot Aliengo, showing consistent mapping performance under lunar gravity and terrain conditions.

cs.RO↗

arXiv2025

RAKOMO: Reachability-Aware K-Order Markov Path Optimization for Quadrupedal Loco-Manipulation

Legged manipulators, such as quadrupeds equipped with robotic arms, require motion planning techniques that account for their complex kinematic constraints in order to perform manipulation tasks both safely and effectively. However, trajectory optimization methods often face challenges due to the hybrid dynamics introduced by contact discontinuities, and tend to neglect leg limitations during planning for computational reasons. In this work, we propose RAKOMO, a path optimization technique that integrates the strengths of K-Order Markov Optimization (KOMO) with a kinematically-aware criterion based on the reachable region defined as reachability margin. We leverage a neural-network to predict the margin and optimize it by incorporating it in the standard KOMO formulation. This approach enables rapid convergence of gradient-based motion planning -- commonly tailored for continuous systems -- while adapting it effectively to legged manipulators, successfully executing loco-manipulation tasks. We benchmark RAKOMO against a baseline KOMO approach through a set of simulations for pick-and-place tasks with the HyQReal quadruped robot equipped with a Kinova Gen3 robotic arm.

cs.RO↗

arXiv2025

PACC: A Passive-Arm Approach for High-Payload Collaborative Carrying with Quadruped Robots Using Model Predictive Control

In this paper, we introduce the concept of using passive arm structures with intrinsic impedance for robot-robot and human-robot collaborative carrying with quadruped robots. The concept is meant for a leader-follower task and takes a minimalist approach that focuses on exploiting the robots' payload capabilities and reducing energy consumption, without compromising the robot locomotion capabilities. We introduce a preliminary arm mechanical design and describe how to use its joint displacements to guide the robot's motion. To control the robot's locomotion, we propose a decentralized Model Predictive Controller that incorporates an approximation of the arm dynamics and the estimation of the external forces from the collaborative carrying. We validate the overall system experimentally by performing both robot-robot and human-robot collaborative carrying on a stair-like obstacle and on rough terrain.

cs.RO↗