뉴스 및 연구

arXiv:2605.18727v2 Announce Type: replace Abstract: Evaluating embodied systems with real dexterous hardware requires more than isolated motor-skill tests: an agent must perceive a changing scene (e.g. a tabletop), choose a context-appropriate action, execute it with a dexterous hand, and leave the scene usable for later decisions. We introduce DexHoldem, a comprehensive real-world benchmark evaluating Texas Hold'em related dexterous manipulations with a ShadowHand. DexHoldem provides 1,470 teleoperated demonstrations across 14 Texas Hold'em manipulation primitives, a standardized physical policy benchmark, and an agentic perception benchmark that tests whether agents can recover the structured game state needed for embodied decision making. On primitive execution, $\pi_{0.5}$ obtains the highest task completion rate ($61.2\%$), while $\pi_{0.5}$ and $\pi_0$ tie on scene-preserving success rate ($47.5\%$). On agentic perception, Opus 5.5 narrowly leads on both strict problem-level accu

arXiv:2605.15510v2 Announce Type: replace Abstract: Robotic hand kinematic design involves combinatorial selection among multiple finger structures whose performance depends on both individual kinematic characteristics and inter-finger interactions. This study presents a potential-dexterity-based quadratic unconstrained binary optimization (QUBO) formulation for combinatorial optimization of robotic hand kinematic structures. Potential dexterity-related metrics were evaluated through kinematic analysis and encoded into a 27-variable QUBO problem incorporating individual finger performance, thumb-to-finger workspace-overlap interactions, one-hot selection constraints, and pairwise structural compatibility. Exhaustive evaluation of the design space established a reference optimum and enabled direct validation of the formulated objective. Discretization sensitivity analysis showed that the selected design remained unchanged across the tested joint-space sampling intervals and at a finer w

arXiv:2602.23934v2 Announce Type: replace Abstract: This paper presents a novel autonomous robotic assembly framework for constructing stable structures without relying on predefined architectural blueprints. Instead of following fixed plans, construction tasks are defined through targets and obstacles, allowing the system to adapt more flexibly to environmental uncertainty and variations during the building process. A reinforcement learning (RL) policy, trained using deep Q-learning with successor features, serves as the decision-making component. As a proof of concept, we evaluate the approach on a benchmark of 15 2D robotic assembly tasks of discrete block construction. Experiments using a real-world closed-loop robotic setup demonstrate the feasibility of the method and its ability to handle construction noise. The results suggest that our framework offers a promising direction for more adaptable and robust robotic construction in real-world environments.

arXiv:2602.11291v3 Announce Type: replace Abstract: World models are becoming central to robotic planning and control by predicting future state transitions. Existing approaches mainly rely on visual, latent, or language prediction, which can be difficult to ground in executable robot actions and prone to compounding errors over long horizons. In contrast, traditional robotic task and motion planning enables structured long-horizon reasoning through compact symbolic representations of world transitions, but typically lacks synchronized visual prediction. We propose Hierarchical World Model (H-WM), which jointly predicts logical and visual state transitions by combining a high-level logical world model with a low-level visual world model. The predicted logical actions and latent visual state transitions are jointly incorporated into Vision-Language-Action (VLA) models as intermediate state guidance for long-horizon task execution. Experiments on three long-horizon benchmarks and real ro

arXiv:2601.20701v2 Announce Type: replace Abstract: Real-time robot control demands fast action generation. Diffusion and flow matching policies for robot control require multi-step sampling, limiting their deployment in real-time scenarios. Natively reducing the sampling steps to one sacrifices representation quality and task performance, creating a trilemma among speed, fidelity, and performance. We present One-Step Generative Policy Optimization (OGPO), a systematic framework to resolve this trilemma. OGPO first pairs a lightweight architecture with the interval velocity principle for distillation-free one-step inference, while representation spreading prevents representation quality degradation. It then performs on-policy reinforcement learning (RL) fine-tuning on this fast, stable policy to break the imitation learning ceiling. Experiments on RoboMimic and OpenAI Gym benchmarks show that OGPO matches or exceeds multi-step baselines while achieving a 5-20 times inference speedup an

arXiv:2510.26142v2 Announce Type: replace Abstract: Trajectory planning for mobile robots remains a major challenge, particularly in cluttered environments, where existing planning algorithms often fail or produce suboptimal paths. To address this issue, we propose an adaptive trajectory refinement algorithm, ART-TEB, comprising two main stages. First, to ensure safety at the path-segment level, a segment-wise conservative collision test is applied, recursively subdividing risky path segments until the collision risk is eliminated. Second, to guarantee pose-level safety, pose correction based on separation direction and line search is applied, ensuring that each pose in the trajectory is collision-free and maximally clear from obstacles. Simulation results demonstrate that the ART-TEB achieves up to 3.12x higher success rates and up to 23.7x faster average planning times than state-of-the-art approaches. Furthermore, real-world experiments confirm that the robot can safely pass through

arXiv:2503.14669v2 Announce Type: replace Abstract: Robot manipulators may start a new task with a tracking error larger than the prescribed tolerance. Conventional barrier controllers generally require the initial error to lie within this tolerance, which prevents their direct use under such conditions. This paper develops a progressive barrier controller that gradually contracts an initial error bound to the required value within a prescribed time. The robot can therefore start outside the final bound, while the direct joint-position error satisfies it after the transition. The closed-form control law combines progressive barrier feedback with an online adaptive torque term based on an actor--critic structure. A Lyapunov analysis establishes bounded closed-loop signals and gives sufficient conditions for satisfaction of the final tracking bound. Two-link simulations consider large initial errors, actuator saturation, dynamic variations, disturbances, and measurement errors. The adapt

arXiv:2503.10118v3 Announce Type: replace Abstract: The sim-to-real gap remains a critical challenge in robotics, hindering the deployment of algorithms trained in simulation to real-world systems. We propose a flexible Real-to-Sim-to-Real (RSR) framework whose central contribution is an information-theoretic cost function that explicitly accounts for sim-to-real discrepancies. This cost balances two objectives, completing the task and steering the policy to collect real-world samples that are maximally informative for improving transfer. It can be integrated seamlessly into existing reinforcement learning algorithms (e.g., PPO, SAC) and ensures a balanced exploration of critical regions in the real domain. The framework treats differentiable simulation as optional: when a differentiable simulator is available, the collected informative data can also be used to tune simulator parameters. We implement the framework with the MuJoCo MJX platform and demonstrate its generality by evaluatin

arXiv:2609.40244v1 Announce Type: cross Abstract: Mobile robots and vehicles carry synchronized multi-camera rigs, yet many streaming 3D foundation models are designed for monocular input, leaving efficient use of rig geometry a challenge. We present StreamRig, a freeze-and-stream framework that builds causal streaming odometry for calibrated rigs on a frozen multi-view 3D foundation model. The frozen front-end jointly perceives the synchronized views using rig calibration. A Rig-Resampler compresses their features, a CausalBridge applies causal attention with a key-value cache, and a lightweight head regresses rig poses. A periodic re-anchoring protocol supports stable pose estimation over long sequences. Only these modules are trained, 74.6M parameters in total, with relative poses as the sole supervision. Our two-stage training strategy combines group relocalization pretraining with causal rig training to transfer the geometric priors of the frozen front-end and the alignment abilit

arXiv:2609.39888v1 Announce Type: cross Abstract: Neural trajectory predictors can reach low prediction error while violating dynamics, actuator limits, or state constraints, especially when controls are unobserved and dynamics are partially specified. We introduce the Markovian Dynamics Enforcer (MaDE), a time-invariant post-hoc operator mapping state-transition proposals onto a learned feasible dynamics manifold, trained on feasible states without ground-truth controls. For each transition it infers a control and recomputes the state through a completion model of known physics plus a learned residual. It then corrects that control by gradient-based inequality reduction, so inequality satisfaction is best-effort within an iteration budget. Since every correction iterate re-enters the completion model, the returned state is dynamically consistent by construction relative to that model and the supplied previous-state anchor. MaDE drives dynamics residuals to essentially zero on fully sp