ニュース＆研究

arXiv:2610.41330v1 Announce Type: new Abstract: Motion planning in clutter is embodiment-specific: the feasible free space and reachable commands depend on each robot's geometry and dynamics. Existing cross-embodiment policies output normalized commands that disregard these differences, or compensate through per-robot training and distillation. World models can in principle pool cross-embodiment data, yet command-conditioned models bind their latent dynamics to a single action space, while pose-conditioned models omit the mapping from commands to body-specific motion. On flat ground, however, embodiment affects the planner's observations only through the camera ego-motion. This letter proposes Hourglass, an hourglass-shaped architecture that factorizes cross-embodiment motion planning at this interface: a body-agnostic world model and each robot's differentiable ego-dynamics, which map commands to that ego-motion, are coupled only through it. The world model is trained on data pooled a

arXiv:2610.41320v1 Announce Type: new Abstract: Recent multimodal foundation models exhibit strong perception, reasoning, and in-context learning capabilities, but translating these capabilities into physical action remains challenging. Through real-world experiments across diverse manipulation tasks, we find that ChatGPT-6-Astra, a state-of-the-art general-purpose reasoner, can serve as a physically actionable, context-responsive manipulation policy, demonstrating effective fine manipulation and spatial reasoning. Building on these findings, we introduce LEXP, a harness for learning from physical experience through context-space adaptation. LEXP combines task-level planning with an experience module that interprets execution evidence, extracts and retains lessons with applicability conditions, and reuses relevant lessons to generate contextual feedback for subsequent actions. Model weights and the action interface remain fixed throughout. Across 12 real-world manipulation tasks, LEXP

arXiv:2610.41275v1 Announce Type: new Abstract: Embodied visual tracking requires a robot to follow a language-specified person under occlusion, look-alike distractors, and ambiguous instructions. We show that improving tracking persistence can simultaneously increase safety violations. Using a compact 0.6B vision-language-action (VLA) tracker with a frozen backbone, we update only the planner head with proximal policy optimization (PPO). PPO improves success and following rates across all three EVT-Bench settings, but increases the collision rate on distracted tracking because episodes that previously ended in target loss can instead end in contact. We introduce TRACE, an execution-time target-state risk gate that preserves the learned policy and intervenes selectively. A lightweight risk head processes 84 cues, including candidate boxes, identity margins, a learned range-and-bearing probe, and the policy's commands. When the calibrated risk exceeds a threshold, TRACE rescales the tra

arXiv:2610.41235v1 Announce Type: new Abstract: Centralized multi-agent path planning with cooperative localization in GPS-denied environments is generally computationally intractable when formulated in full belief space due to tightly coupled uncertainty dynamics. Prior work typically approximates belief dynamics or restricts inter-agent coupling. This work examines a multi-agent planning problem in which a primary agent must reach a goal in minimum time subject to a chance-constrained collision-avoidance requirement and a terminal bound on localization uncertainty. We obtain a tractable surrogate of the underlying belief-space problem by optimizing over the mean-state trajectories of all agents and propagating covariance for terminal constraint evaluation only. We propose CONSORT (Cooperative Navigation with Support-Optimized Routing and Trajectories), a discrete-to-continuous pipeline combining hexagonal-grid joint search with spline-based trajectory refinement. Across 50 randomized

arXiv:2610.41216v1 Announce Type: new Abstract: Robot agents built on vision-language models depend on both the planner and its harness: the tools, feedback, memory, and recovery logic around it. Can improvements to a large teacher's harness be carried into a compact agent with lower inference compute cost? We present TinyAstral, a two-stage approach. We first freeze an open-weight teacher and improve its harness through recursive self-improvement (RSI): frontier LLM agents read the teacher's trajectories and propose harness edits, and each edit is accepted only if it passes a paired statistical test. We then freeze the final harness, collect teacher trajectories with reasoning and failure recovery, and distill a small student that replaces the teacher in the same harness. On RoboTwin, harness evolution raises the teacher's success from 33.4% to 63.6%; the final 4B student reaches 62.8%. Across four harness versions, distilled 4B students score close to their corresponding teachers, an

arXiv:2610.41214v1 Announce Type: new Abstract: Multi-robot systems are increasingly deployed as heterogeneous fleets with diverse kinematic and actuation capabilities. While graph control barrier function (GCBF) methods provide a scalable framework for decentralized safe control, existing approaches typically assume homogeneous robot dynamics. Methods that directly encode robot classes can distinguish known types, but do not represent state-dependent motion capabilities or naturally transfer to unseen robot models. To address this challenge, we propose the Capability-Conditioned Heterogeneous Graph Control Barrier Function (CC-HGCBF), defined in a shared second-order reference-point operational space. This formulation provides a common geometric representation of safety across heterogeneous robots while retaining their model-specific dynamics and admissible control constraints. To enable zero-shot transfer, we introduce the Radial Capability Field (RCF), a compact, fixed-dimensional r

arXiv:2610.41188v1 Announce Type: new Abstract: Safe navigation in heterogeneous multi-robot teams faces two coupled challenges: differences in robots' braking and turning capabilities limit how each robot can respond to an approaching neighbor, while simultaneous interactions require a single control action to satisfy multiple collision-avoidance constraints. Together, these challenges can make local safety constraints incompatible with a robot's admissible controls, leaving a decentralized safety controller without a feasible action. We propose a Capability-Aware Backup-Recoverability Filter (CA-BRF), based on the insight that navigation actions should preserve access to a certified recovery maneuver. First, we construct a backup-recoverable set in a shared Cartesian operational space by evaluating whether prescribed backup maneuvers maintain pairwise safety until the robots come to rest. Each maneuver respects the corresponding robot's kinematic and actuation constraints, making the

arXiv:2610.41185v1 Announce Type: new Abstract: End-to-end (E2E) autonomous driving models achieve strong planning performance but remain largely opaque, limiting observability and trust. We apply mechanistic interpretability to E2E driving by training Sparse Autoencoders (SAEs) on the internal activations of a simple, single-camera end-to-end path planning model. Unsupervised clustering of SAE latents reveals interpretable structure aligned with route intent, vehicle speed, and braking behavior, recovering route intent with 98.3% accuracy without supervision. Through paired counterfactual image edits, we further show that specific SAE features are visually grounded and systematically respond to scene changes such as traffic light state and stop sign presence. Steering individual SAE latents produces predictable, monotonic changes in planner outputs including lateral displacement, curvature, and braking. Finally, we demonstrate a practical application of these representations: we build

arXiv:2610.41053v1 Announce Type: new Abstract: Recent years have witnessed increased momentum towards memory-safe programming languages for safety-critical systems. Memory-safe languages can eliminate entire classes of common software errors that have historically required extensive analysis and expenditure to avoid. Nevertheless, memory-safe language adoption in aerospace remains fragmented and incomplete, particularly for unmanned aerial vehicles (UAVs). Recent Rust UAV projects either run only in simulation, are closed source, or have not been integrated with an established autonomy ecosystem and tested against existing C/C++ firmware. To address this gap, we present Veloxity: an open-source embedded UAV autopilot written in the memory-safe Rust programming language. Veloxity lowers the barrier to using memory-safe languages within modern ecosystems by providing a minimal, modular foundation for autonomous UAVs. We compare Veloxity to equivalent autopilots written in C++ in simulat

arXiv:2610.41028v1 Announce Type: new Abstract: Robot-free human demonstrations offer a promising route to scaling dexterous manipulation data. The key challenge is to preserve the demonstrator's natural dexterity while producing contact-grounded observations and actions that transfer to robot hands. We present Contact Imitation Interface (CIMI), a framework with contact-grounded demonstration and learning interfaces. The demonstration interface combines shared contact geometry and fingertip RGB views with contact-alignment-first retargeting and reprojection of wider scene views. The learning interface introduces a hierarchical policy in which a contact-surface planner predicts relative contact-surface trajectories and a controller coordinates wrist and finger commands to improve contact alignment. CIMI combines contact-grounded observations, retargeting, and action representations, improving mean final-stage success by at least 50 percentage points over baseline policies across six ha

arXiv:2610.41009v1 Announce Type: new Abstract: Vision-language-action models (VLAs) map language instructions and visual observations to robot actions, showing strong capabilities across robotic manipulation tasks. However, in our probing experiment, we find that action representations near the action output do not fully preserve task-relevant visual grounding. Motivated by this late-layer grounding gap, we introduce STAR, for STeering Action Representations, a training-free inference method for pre-trained VLAs. STAR applies an exponential moving average over late-layer residual updates, allowing later action representations to inherit object-aware update directions from earlier layers without model updates or architectural changes. It further uses recent proprioceptive states to adapt the mixing strength, using conservative mixing when execution dynamics change rapidly. Experiments on robot manipulation benchmarks show that STAR improves late-layer object awareness and increases ave

arXiv:2610.40922v1 Announce Type: new Abstract: End-to-end policies, from task-specific controllers to robot foundation models, learn complex behavior from raw observations but are not guaranteed to remain safe, so they are often wrapped in a safety filter. The resulting system operates under two domains: the filter's operational design domain (ODD), where its assumptions hold and safety is certified, and the policy's competence domain, where it is reliable. Filters typically fix their conservativeness regardless of the policy's competence, trading performance inside the ODD against margin beyond it. We ask whether the policy's confidence should set that level, and propose a filter-agnostic pipeline that modulates it at runtime: inside the ODD, where every setting is safe, confidence buys performance; beyond it, low confidence adds margin, which helps when it flags where the filter's model fails. Preliminary F1TENTH simulations, where this holds, show the modulated filter combining the

arXiv:2610.40902v1 Announce Type: new Abstract: Flapping-wing simulation must reconcile high-frequency wingbeat aerodynamics with the slower dynamics relevant to flight control and reinforcement learning. We present FlapSim, a modular open-source MuJoCo simulation platform that combines robot descriptions, wing geometry, and composable aerodynamic models in a deterministic simulation pipeline. Across $48$ held-out scenarios spanning frequency, gimbal, wind, frequency-step, gust, and cross-robot flat-plate regimes, we compare five reduced-order configurations operating at $1\,\text{ms}$ with a $50\,\mu\text{s}$ quasi-steady blade-element reference. The built-in hover cycle-averaged baseline achieves a $67.51\times$ CPU speedup but only $0.6593$ mean wrench fidelity. Combining multirate $21$-substep clock synchronization, three-dimensional linear/quadratic inflow and gust coupling, analytical blade-element geometry scaling, and phase-locked $1f\text{--}6f$ harmonics raises mean fidelity

arXiv:2610.40828v1 Announce Type: new Abstract: Imitation learning policies are intended to make robotic manipulation robust to variations in object pose, appearance, and environment that challenge pre-programmed motion. However, real-world performance remains difficult to predict, as conventional optimization signals such as validation loss show no consistent relationship with robot success. To identify the design choices that govern real-world performance, 281 imitation-learning policies spanning seven policy families were evaluated in approximately 1,700 real-robot trials of RJ45 connector grasping and insertion using a five-fingered robotic hand. Controlled ablations showed that observation and action representations dominated conventional optimizer and sampling choices. For point-cloud policies, region-of-interest cropping, representation in the robot-base frame, and normalization within the point encoder each improved success by 70-100 percentage points. For the vision-language-a

arXiv:2610.40799v1 Announce Type: new Abstract: Across coding, web, and tool-use tasks, language-model agents are now routinely trained with the same recipe: supervised fine-tuning (SFT) to learn the interface, then multi-turn reinforcement learning (RL) to master it. Code-as-policy (CaP) casts robot manipulation in the same form, as the policy writes code against robot APIs and reads back the results, so the same recipe should apply. Yet existing CaP agents sidestep training: they either prompt closed frontier models or delegate manipulation to pre-trained vision-language-action models (VLAs) and high-level skills, while small open vision-language models (VLMs) fail outright. We present CaPtain, an open post-training pipeline that turns a small open VLM into a capable CaP agent, with no VLA, frontier model, or privileged state at test time. CaPtain trains and deploys the policy inside CaPtain-Harness, which exposes low-level perception, geometry, and guarded-motion primitives across s