Today's largest cluster—five VLA and foundation-model papers—converges on a theme distinct from the field's usual march toward bigger multimodal pretraining: adding a verification or self-monitoring layer on top of otherwise black-box action generation, rather than trusting a single forward pass. EmbodiedSkills treats every skill decision as an execution proposal that must pass a prerequisite check before running and an outcome check afterward; VerNav replaces expensive per-step autoregressive LLM reasoning in vision-and-language navigation with a lightweight verifier that only escalates to a full generator when uncertain, cutting decision latency more than 10× on Room-to-Room; and Knowing When to Stop reads a VLA's own cross-attention entropy to decide, training-free, when an action chunk has run out of grounding in the current observation. DroneCATS makes the underlying motivation explicit: evaluating MLLMs dropped directly into a drone control loop, it finds that small open models often navigate as well as frontier models but lose episodes by declaring arrival prematurely or never declaring it at all — spatial competence is not the bottleneck, protocol discipline is. REFACTOR-VLA and Facet-0 push structure earlier in the pipeline instead: the former learns a typed skill library gated by description-length and return-preservation tests (and finds that a bigger 430M-parameter world model actually hurt clustering versus a better training objective), the latter builds an explicit action-wrench critic so a manipulation policy can distinguish superficially similar motions by their contact outcome, reaching 82% success on sub-millimeter assembly versus 15% for the strongest baseline.
A parallel cluster of five papers pursues the same instinct from the control-theory side: keep a model-based or classical controller's guarantees while letting learning handle what it can't specify in advance. ProxPI proves, not just demonstrates, that centering MPPI's sampling distribution on an out-of-distribution learned policy traps the optimizer near a bad solution regardless of rollout budget, and fixes it by injecting the policy only as a soft proximity cost around nominal-centered sampling. SG-RL takes the mirror-image approach for RL-based MPC weight adaptation, keeping sampled closed-loop return as the objective while injecting bounded solver-derived gradients as auxiliary guidance, reaching PPO's best return with up to 70.6% fewer samples on real racing platforms with intentional model mismatch. Dual Process Motion Planning states the philosophy outright, borrowing "Thinking Fast and Slow" to pair a symbolic System-2 solver with experience-driven System-1 modules under a metacognitive controller. Even the more classical entries fit the pattern: the PID regulatability paper proves a long-standing engineering heuristic — tuning gains channel-by-channel in coupled MIMO systems — is theoretically justified rather than just convenient, while the Rao-Blackwellized POMDP paper narrows its efficiency gains to problems with tractable sufficient statistics rather than claiming a general fix. None of these are the flashiest results in the batch, but together they suggest the sim-to-real, learning-versus-classical-control debate is settling into hybrid designs with formal guarantees on the learned component's failure modes, not just empirical ablations.
A third theme is robustness to missing information, expressed at very different scales. SCARAB and the connectivity-aware exploration paper both design multi-robot coordination that assumes communication will fail — the former dropping GPS and RF entirely for a river-bridging swarm, the latter extending exploration graphs specifically to keep area partitioning stable as the communication graph itself becomes unstable. HitMem tackles the same problem for a single embodied agent's memory, adding temporal decay and a two-stage retrieval escalation so that objects moved by unobserved events don't force costly re-exploration. At the hardware level, the compact CVT robotic finger and MONORIGAMI folding actuator both push adaptivity into the mechanism itself rather than the controller: the finger's passive transmission amplifies grip force up to 4.19× purely from wire tension with no added actuators or sensors, and MONORIGAMI converts a single 3D-printed part into three different multi-DoF systems via programmed stiffness anisotropy alone. One caveat worth flagging: Semantic Scholar's author lookup returned no matches for any of today's 30 papers, so the h-index badges below are uninformative today — the ordering reflects arXiv submission order, not author seniority.
Vision-language-action and foundation-model papers that add verification, contact-awareness, or skill-abstraction layers on top of action generation.
5 papers
Contact-rich manipulation: temporal robustness under speed variation, reconfigurable insertion benchmarking, and RL-driven non-prehensile throwing.
3 papers
Active perception, correspondence-free registration, dynamic 3D memory, low-latency VLN, and camera-controlled generation grounded in real-world geometry.
6 papers
Coordination under degraded or absent communication and GPS: UAV formation, communication-free swarm bridging, and exploration graph stability.
3 papers
Hybrid model-based/learned control and planning: MPC-RL fusion, Rao-Blackwellized POMDPs, neuro-symbolic planning, and classical PID theory.
5 papers
Vehicle-level context modeling, noise-aware speed control, and lane-change intention recognition for ADAS and automated driving.
3 papers
Mechanism-level adaptivity: a passive-CVT robotic finger, an origami-inspired folding actuator, and low-cost embedded sensor fusion.
3 papers
Runtime-editable loco-manipulation behavior architecture and diffusion-based language-conditioned whole-body humanoid control.
2 papers
Real-world robotic assembly at sub-millimeter tolerances demands spatial precision, compliant interaction, and robustness to contact failures. We present Facet-0, a robotic foundation model that predicts and values the contact consequences of its actions. Facet-0 unifies multimodal representation learning and reinforcement learning (RL) post-training around a joint action-wrench proposal: a causal wrench history is aligned with vision-language semantics and kinematic state, and flow matching generates each action chunk together with the future wrist-wrench profile it is expected to induce. Deployment rollouts train a distributional Action-Wrench Critic to distinguish motions with similar task progress but different contact outcomes, while phase-aware rewards and contact-selective credit concentrate policy improvement on decisive interactions. To accommodate part-specific dynamics, a lightweight bounded actor reuses the frozen representation for on-robot adaptation; RL remains defined over executable Cartesian actions, while an auxiliary wrench head preserves predictive, non-commanded action-contact coupling. Trained on ManuFacet-1K, a 1,000-hour force-synchronized corpus spanning three embodiments and multiple manufacturing cells, the bounded task-adapted system reaches 82% mean success on five sub-millimeter computer-assembly tasks, compared with 15% for the strongest baseline, with 0.5 mm placement accuracy and 50 ms command latency.
Multimodal Large Language Models (MLLMs) are strong perceivers of images and video. We ask how far that reach extends into acting: dropping an MLLM directly into a drone's control loop, with its entire action space declared solely in the prompt. Recent systems approach this setting but increasingly narrow the model's decision-making. We widen it back. We introduce DroneCATS-Agent, an architecture where the MLLM is a swappable component, and DroneCATS, a benchmark treating the model as the independent variable. Beyond merely flying toward a pixel, our agent entrusts the model to yaw and search, deliberate when unsure, and self-declare arrival—all without fine-tuning or function-calling schemas. Evaluating frontier and open models across four core capabilities—approaching a visible target, tracking a moving one, searching outside the initial view, and commanding a multi-drone fleet—reveals that even the simplest embodied settings are far from solved. Crucially, to identify what breaks first at the edge, our roster scales down to 2B parameters. The findings expose a stark paradox: it is not the flying that fails. Small open models often navigate into the success radius more reliably than frontier models, yet lose the episode by declaring arrival prematurely or not at all. Multi-drone commanding amplifies this divide, with small models failing by blindly copying a single coordinate across distinct views. Viewed as vision-language-action agents, the models' spatial perception holds up, but their action protocol does not. What separates a deployable edge model from a frontier model is not navigation, but the discipline to sustain a declared protocol and emit the correct terminating action. The open problem is closing this gap at onboard compute costs—yielding a fast model that plans persistently and knows exactly when it is done—and DroneCATS is built to measure that distance.
Vision-language-action (VLA) models map visual observations and language instructions directly to robot actions, but long-horizon tasks require more than action prediction. An agent must coordinate perception, planning, execution, progress verification, and recovery as the physical state evolves. An action prediction or a model-generated skill decision does not, by itself, guarantee that the proposed operation is valid in the current state or that its outcome will be verified. We propose EmbodiedSkills, a unified framework that treats each skill decision as an execution proposal: the runtime checks its prerequisites before execution and verifies the outcome afterward. A shared executable-skill interface connects high-level skill selection, bounded low-level VLA execution, and post-action verification within a single agent loop. Because this interface remains fixed, low-level VLA policies can be replaced or adapted without changing the agent loop. The interface also records planning, execution, verification, and recovery events as structured trajectories, which provide supervision for individual components and can support optional online adaptation when interactive feedback is available. We instantiate EmbodiedSkills with Qwen3-VL and OpenPI/pi0.5 on RoboTwin 2.0 and LIBERO. Task-adapted low-level VLA policies achieve an average success rate of 86.20% across 50 RoboTwin 2.0 tasks and 97.40% across the four LIBERO suites. These results establish the execution performance of the task-adapted low-level VLA policies used in EmbodiedSkills. On four memory-dependent RMBench tasks, the same task-adapted execution approach achieves 12.5% average success. The framework provides a trainable and inspectable agent layer for turning these policies into closed-loop embodied systems.
Most vision-language-action (VLA) models — OpenVLA, π0, RT-2, RDT-1B — are monolithic: they emit raw motor commands or short action chunks without organizing behavior into reusable abstractions, so they degrade on long-horizon tasks and resist interpretation. Existing skill-discovery methods sidestep the core question of when two action sequences are behaviorally equivalent, either clustering contrastive embeddings or delegating the judgment to a language model uncalibrated to the robot's dynamics. We introduce REFACTOR-VLA, a wake/sleep system for learning reusable skills. Its sleep phase clusters motor-program fragments under a Behavioral-Equivalence Kernel (BEK) computed from rollouts of a learned latent world model M_φ; its wake phase emits typed lambda terms over a Hindley—Milner-inspired vocabulary, consumed by a library-conditioned rectified-flow action decoder. Abstractions are admitted only if they pass Minimum Description Length and return-preservation gates. On LIBERO we report two findings. First, enlarging the world model from 188M to 430M parameters worsened performance on 4 of 4 suites, so capacity alone does not help. Second, the training objective matters far more: adding an auxiliary supervised contrastive (InfoNCE) loss during world-model warmup substantially improves sleep-phase clustering, giving Normalized Mutual Information at n=3 seeds of 0.462 \pm 0.021 (object), 0.867 \pm 0.025 (spatial), 0.915 \pm 0.013 (goal) and 0.754 \pm 0.010 (LIBERO-10), and beating the strongest published baseline on all 4 suites by a mean Δ=+0.184. Across providers (n=12) the 95% bootstrap confidence interval for mean pairwise NMI is [0.683, 0.729] (mean 0.705). The sleep phase also yields the first real-LIBERO task-language library: the decoder uses 2 of 3 admitted abstractions and rewrites all 256 sampled demonstrations.
Action chunking is a standard execution strategy in modern Vision-Language-Action (VLA) frameworks, but fixed execution horizons impose a trade-off between efficiency and accuracy. Short chunks require frequent inference and may cause oscillatory behavior, whereas long chunks can become misaligned with newly observed states. We address this limitation with an adaptive action chunking approach based on internal cross-attention dynamics in the action expert. We observe that, as the prediction horizon extends, action-to-observation cross-attention becomes increasingly dispersed and its entropy rises toward a plateau. This pattern is associated with higher action prediction error and provides an online signal that the current observation offers limited grounding for further open-loop execution. Based on this observation, we introduce a training-free truncation mechanism that detects sustained high-entropy plateaus and dynamically selects the execution horizon during inference. The method uses attention weights already computed by the policy and introduces negligible additional overhead. Evaluations on π_{0.5} and X-VLA across RoboTwin 2.0, LIBERO, and three real-world manipulation tasks show improved average task success over fixed-horizon and adaptive chunking baselines, while preserving efficient closed-loop control. These results show that cross-attention dynamics can provide a practical internal signal for adaptive action execution in VLAs.
Dexterous manipulation policies learned by imitation are typically evaluated for robustness to variation in scenes, objects, or instructions, but their performance across task execution speeds is less often examined. This leaves open how much temporal robustness a learner retains relative to the expert it imitates. We compare an expert and learner under the same task conditions, initial-condition draws, and speedup factors. We instantiate the evaluation in ParcelStow, a contact-rich task in which the robot acquires, reorients, and inserts a parcel. The demonstrations span the speedup range for the manipulation phases after parcel acquisition. A scripted expert and an Action Chunking with Transformers (ACT) policy trained from the expert's demonstrations both achieve 100 percent task success at nominal speed. Their success rates diverge within the demonstrated range: at its maximum, expert success is 84 percent and ACT success is 53 percent. Two ACT policies with different parameter initializations show similar degradation, decreasing by 34 and 48 percentage points from nominal speed to the maximum demonstrated speed, compared with 16 points for the expert. Stage-level analysis shows that 35 of ACT's 47 failures at the maximum demonstrated speed are insertion misalignments. Under the relative-motion handoff, every ACT acquisition retains the parcel through reorientation and transfer in free space, but only 64 percent complete the overall task, compared with 95 percent after expert acquisition. Across all evaluated policies and speeds, none of the 414 acquisitions without force closure completes the task. Equal nominal task success therefore does not imply preservation of expert performance across execution speeds. Code, data, and evaluation scripts are available at https://github.com/coenwerem/parcelstow.
High-precision insertion remains a fundamental challenge in robotic manipulation due to the strict alignment requirements and contact-rich interactions involved. Although peg-in-hole tasks are widely used for evaluation, existing bench- marks often rely on fixed task configurations, limiting their ability to assess robustness and generalization across different insertion scenarios. This paper introduces a reconfigurable peg-in-hole benchmark designed to evaluate task generalization in high-precision insertion. The benchmark consists of a set of fully 3D-printable modular components, including multiple peg geometries, tolerance levels, and configurable base structures that can be combined to generate a large variety of insertion and assembly tasks. By varying object layouts, orientations, and task structures while maintaining controlled physical conditions, the benchmark enables systematic evaluation of adaptation to unseen scenarios. To support reproducibility, we additionally provide a scenario generation tool capable of producing standardized task configurations and machine-readable task descriptions. The scenario generation tool and the STL files of the benchmark pieces are available through the project repository: https://github.com/aistairc/peg-in-bench.
Robotic throwing enables fast object transport and extends a robot's reachable workspace beyond traditional pick-and-place. While prehensile (grasp-based) throwing works well for graspable items, non-prehensile (grasp-free) throwing is better suited for large, heavy, and/or deformable objects. Existing approaches rely on model-based optimization with simplified contact models (e.g., dynamic grasping) and low-dimensional trajectory parameterizations, which limit solution quality and reachable workspace. We propose a reinforcement learning approach that additionally leverages sliding and rolling contact modes and directly optimizes joint-space trajectories without analytical contact models or custom parameterizations. The Markov Decision Process (MDP) is formulated as a dynamical system that evolves the robot's joint state conditioned on the throwing target, object model, and initial configuration. Joint-jerk trajectories are planned offline at a low control rate and upsampled into smooth, high-rate velocity commands for deployment. For sim-to-real transfer, we minimize the robot-dynamics gap through minimum-jerk system identification and train uncertainty-aware policies to mitigate object-modeling errors, particularly sensitivity to dynamic friction. In simulation, the policy achieves 99% success across thousands of configurations and generalizes to unseen objects. Sensitivity analysis shows robustness to mass uncertainty but high sensitivity to dynamic friction, consistent with the sliding-based release mechanism. Deployed zero-shot on a UR5e operating near its physical limits (5 m/s end-effector velocity), our method throws diverse objects including heavy (790 g) and large (20x20x28 cm) items to targets up to 350 cm distance or 180 cm elevation, achieving a 97% real-world success rate.
We present SG-AMP, integrating robust depth completion with input-conditioned uncertainty, persistent panoptic mapping, plant scene-graph reasoning, and semantics-aware active view-motion planning. Beyond inspecting uncertain observed regions, the scene graph explicitly hypothesizes unobserved pepper—peduncle attachments and directs close-range sensing toward them. Candidate views are selected according to expected information gain, while class-dependent motion costs distinguish protected peppers, peduncles, and stems from conditionally traversable foliage. On pepper data, the perception network achieves 55.27% semantic mIoU, 38.67% PQ, and 40.62\,\mathrm{mm} depth RMSE, while input-conditioned uncertainty improves NYUv2 NLL from -1.6518 to -1.6925 and AUSE from 0.0102 to 0.0087.
Long-duration outdoor coverage with autonomous platforms remains challenging beyond classical planning: deployments face localization drift in open spaces, obstacles in cluttered sites, controller feasibility in turn-heavy maneuvers, and persistent autonomy with energy management. We propose a unified ROS 2 architecture for outdoor coverage that combines coverage planning, robust localization, and Nav2-based execution. A dual-antenna RTK-GNSS fused in an EKF keeps the robot pose, both position and heading, accurate across long missions; three controller-aware refinements are added to a mature coverage planner; a Behavior-Tree mission manager coordinates multi-goal execution, layered recovery, cost-aware goal management, and autonomous docking for return-to-charge. We validate the stack through simulation and real-world trials across multiple outdoor areas with varying geometries and obstacle densities. Overall, these results show that the proposed stack can reliably complete outdoor coverage missions across varied areas, sweeping 93.1% to 96.1% of the planned coverage area.
In camera-controlled video generation, geometry-aware positional encodings condition tokens on camera extrinsics and per-token viewing rays. Existing schemes, however, have a scale-dependent failure mode on real-world metric camera trajectories: homogeneous projective encodings cause attention logits and feature norms to grow unbounded with physical translation baselines. We propose MeRoPE (Metric Rotary Position Embedding), a norm-preserving relative camera encoding for attention. MeRoPE encodes relative orientations between calibrated viewing rays with orthogonal rotation blocks, maps raw metric displacements into multi-frequency rotary phases, and adds a disparity-anchored correspondence prior along the epipolar arc. This design strictly preserves feature norms, bounds pre-softmax attention logits regardless of the physical translation scale, and maintains exact invariance to global rigid coordinate changes. Across nuScenes and PanShot, which cover large-baseline trajectories and diverse camera optics, respectively, MeRoPE achieves stronger camera control than prior encodings, with the best consistency between generated camera motion and conditioning poses in both rotation and translation. Code will be made publicly available.
Robotic processing of irregular steel scrap requires dense 3-D measurement to replace manual visual assessment in hazardous cutting workcells. The reconstructed map is used to estimate piece dimensions, boundary geometry, feasible preheating and cutting regions, and collision-aware torch paths. The reconstruction errors therefore propagate directly to downstream measurement and planning. Existing multi-view registration methods commonly rely on feature extraction and data association to establish correspondences between views. In workcells with smooth metallic surfaces, repeated structures, occlusions, and partial overlaps, however, wrong correspondences may be established, leading to inaccurate pose estimation and distorted reconstruction. This paper presents an adaptive layered depth-map-guided bundle adjustment framework for correspondence-free multi-view point cloud registration. The scene is represented by a global 2.5-D grid, where each cell can adaptively maintain multiple depth hypotheses. Raw depth observations are directly projected into the global map to form depth constraints without explicit feature correspondences. At grid cells where multiple surfaces produce conflicting depths, a softmax-based layer assignment links each observation to compatible depth hypotheses. The resulting nonlinear least-squares formulation jointly refines sensor poses and the layered depth map, with correspondences implicitly induced by the depth-map representation and projection model. Experiments on self-collected industrial datasets show that the proposed method achieves consistently competitive reconstruction accuracy while maintaining robustness and low computational cost in challenging industrial scenarios. We release the open-source code implementation at: https://github.com/YiranZhou-Robotics/ADM-BA.git
Executing long-term tasks in dynamic environments requires embodied agents to maintain robust and adaptive 3D scene representations. However, most existing 3D memory frameworks rely on static world assumptions. When objects are displaced by human activities or unobserved events, agents encounter memory-observation conflicts and often require costly geometric recomputations or inefficient global re-exploration. To address this, we propose HitMem, a hierarchical temporal 3D memory framework with a multi-modal context-aware retrieval mechanism. Through continuous perception, HitMem unifies semantic and spatial information into a lightweight topological graph that captures support relationships, while a temporal decay mechanism dynamically regulates memory activeness to mitigate the impact of stale representations. In addition, the multi-modal context-aware retrieval mechanism defaults to filtering candidates using integrated semantic, spatial, and temporal memory features, and activates a specialized two-stage retrieval process when object displacement is detected. This process combines spatial constraints inferred from external agent trajectories with semantic common sense grounded in class affinities, efficiently identifying high-probability candidate regions. Extensive evaluations on our constructed Dyna-THOR benchmark demonstrate that HitMem significantly improves object relocation accuracy, reduces exploration costs, and enhances task execution performance in dynamic environments.
Vision-and-Language Navigation (VLN) requires an agent to navigate through unseen 3D environments according to natural-language instructions. Explicit reasoning can improve instruction understanding and semantic grounding, but autoregressive generation at every step accumulates large decision-stage latency over multi-step navigation. We propose VerNav, a verifier-first framework for low-latency LLM-based VLN. The verifier reduces decision-stage latency by replacing per-step autoregressive generation with batched action verification, while an entropy-based adaptive generator is invoked only for uncertain decisions to produce compact state evidence. To further improve navigation performance with the verifier, we introduce a two-stage alignment scheme: (i) VPO improves local action-preference alignment in static verifier training, and (ii) step-level reinforcement fine-tuning provides dense progress rewards over multi-step navigation rollouts during dynamic task execution. Experiments on the Room-to-Room (R2R) benchmark show that the verifier-only decision path of VerNav achieves competitive navigation performance among representative LLM-based VLN agents while reducing average decision-stage LLM latency per step by more than 10\times compared with autoregressive methods.
Cooperation in multi-UAV systems requires reliable relative perception so that follower vehicles can maintain formation and continue their mission safely even when absolute positioning sensors degrade or fail. This paper presents a vision-based cooperative formation framework running on a follower UAV that uses a front-facing RGB-D camera to detect, track, and localize a leader UAV in real-time. A lightweight YOLO-based detector is trained on a dedicated drone dataset and deployed onboard to predict leader bounding boxes, which are then fused with depth information via a pinhole camera model to estimate the leader's relative pose. These estimates provide a leader-follower position controller and can also be used as a backup when GPS or external localization is unavailable. This framework is implemented as a set of ROS nodes and evaluated in a physics-based multi-UAV simulation built on XTDrone, with sensor noise and communication dropouts. We evaluate detection accuracy, runtime, and formation-keeping error under nominal conditions and under simulated failures of the positioning sensors. The results show that the proposed framework maintains stable leader-follower formations with reasonable computational cost and provides a practical basis for extending vision-based cooperative formation control to real-world multi-UAV systems.
We describe SCARAB—Swarm-Capable Autonomous Robotic Aquatic Bridging. Using distributed swarm control and multi-model sensing of agents and docking targets, agents can localize themselves, join into formations and proceed to desired target locations. Our technologies would eventually allow the Army to perform unpredictable, dispersed river crossings, enhance crew survivability, and minimize the logistics footprint compared to the current Improved Ribbon Bridge. Our methods operate without GPS or RF communications, though these can be used in non-contested environments (e.g., civilian disaster relief for flooding, etc.). We demonstrate our system via physics-engine-based simulation of several agents using unmanned surface vehicles (USVs) in the presence of currents and wind.
Exploring unknown environments with multiple UAVs requires coordination under intermittent communication, making decentralized operation a baseline assumption. We propose, within a decentralized framework, a novel exploration graph extension strategy based on frontier connectivity to extend exploration plans and maintain area partitioning among agents stable and robust to disconnections and changes in spatial layout. The proposed extension method is applied to two state-of-the-art area partitioning methods and evaluated in simulation. Experiments show improved performance over existing graph extension approaches with higher exploration efficiency under low communication rate.
Online planning under uncertainty remains a fundamental challenge for robotic systems operating in partially observable environments with high-dimensional state spaces. While sampling-based POMDP solvers enable approximate decision-making in large or continuous domains, their performance degrades as belief dimensionality increases due to the high variance inherent in Monte Carlo-based estimation. In this work, we extend the Rao-Blackwellized online POMDP (RB-POMDP) framework to improve its generalizability in high-dimensional settings through hybrid continuous-discrete belief representations. By analytically propagating uncertainty associated with marginalized state components during tree-based planning, the proposed approach reduces sampling-induced variance in value estimation. We demonstrate the effectiveness of this framework in a robotic search-and-rescue task by integrating it with FastSLAM 2.0. Experimental results show that the proposed planner achieves higher cumulative rewards using significantly fewer particles and planning simulations than purely sampling-based methods under equivalent computational budgets. These results suggest that structured high-dimensional robotic problems admitting tractable sufficient statistics can be effectively leveraged within the RB-POMDP framework for computationally feasible online decision-making.
Robotic systems are deeply embedded in both industry and everyday life, where they are expected to act with speed, precision, and reliability. Classical control and planning methods have long delivered strong guarantees, but often at the cost of computational efficiency and adaptability. More recently, learning-based approaches have shown promise in overcoming these limitations, enabling agents to leverage experience to accelerate decision-making and address previously intractable problems. In this work, we bridge these two approaches through a neuro-symbolic perspective on nonlinear motion planning. Inspired by the Thinking Fast and Slow paradigm, we introduce a dual-process architecture that combines the strengths of robust reasoning and learning. Our framework integrates state-of-the-art symbolic solvers as a ``System-2'' component with experience-driven ``System-1'' modules. A metacognitive controller dynamically orchestrates their interaction, selecting when to rely on fast intuition versus slower, more precise reasoning. By evaluating the framework across diverse nonlinear benchmark environments, we demonstrate that this architecture yields consistent gains in planning efficiency, accuracy, and generalization, while promoting reuse across tasks. The results suggest that tightly coupling learning with structured reasoning offers a scalable path toward more capable and adaptive robotic systems.
This paper studies a class of uncertain multi-input multi-output (MIMO) nonlinear systems using extended PID (EPID) control. We focus on systems possessing a well-defined vector relative degree whose components may vary across channels, a setting that received limited attention in the existing literature on PID-type control. We develop a channel-wise EPID controller, where each control input is constructed from the proportional, integral, and derivative terms of the corresponding tracking error, and the highest derivative order is selected according to the relative degree of that channel. Under suitable growth conditions on the unknown nonlinearities, we construct an admissible set of EPID parameters specified by prescribed initial-state bounds, the uncertainties and the reference signal. We show that any parameter choice from this set guarantees semiglobal stability of the closed-loop system and achieves the desired setpoint regulation. These results indicate that the EPID parameters can be designed independently for each channel, justifying the common engineering practice of channel-by-channel tuning in strongly coupled and uncertain MIMO systems, including flight control systems.
In Model Predictive Control (MPC), cost-function weights shape closed-loop behavior, yet changing conditions often make fixed parametrizations suboptimal and motivate context-dependent online adaptation. Learning such policies is difficult because behavior depends implicitly on numerical MPC solutions, producing nonlinear, potentially nonsmooth, long-horizon dependencies on policy parameters. This creates a bias-variance tradeoff: Reinforcement Learning (RL) optimizes realized closed-loop return from environment samples but is sample-inefficient, whereas Gradient-Based Policy Learning (GB-PL) uses low-variance solver gradients from differentiable MPC to optimize surrogate losses on predicted trajectories but can be biased under model mismatch. We propose Solver-Gradient Guided Reinforcement Learning (SG-RL), a solver-sensitivity augmentation for RL-based online MPC cost-weight adaptation. SG-RL keeps sampled closed-loop return as the objective and uses bounded solver-derived gradients as auxiliary guidance to improve stability and sample efficiency. We instantiate SG-RL in Proximal Policy Optimization (PPO) with four modular algorithms that inject solver-gradient guidance into actor-update scaling, policy loss, advantage estimation, and value-function learning. On two full-scale autonomous racing platforms with intentional model mismatch, SG-RL reaches PPO's best closed-loop return with up to 70.6% fewer samples, outperforms GB-PL baselines by at least 54% in closed-loop return, and generalizes zero-shot to unseen environments.
Combining learned policies with model predictive control can leverage learned task priors while retaining online adaptation to new objectives and constraints, but performance degrades when the policy is out of distribution. In policy-guided model predictive path integral (MPPI) control, a policy-centered warm-start approach centers the sampling distribution on the policy output. When the prior is mismatched, centering the sampling distribution on the policy output restricts exploration around an unsuitable solution and prevents recovery toward the task optimum. We propose Proximal Prior Injection (ProxPI), which retains nominal-centered MPPI sampling and incorporates the policy through a soft proximity cost. This matches the in-distribution performance of existing prior-injection schemes while enabling the optimizer to escape an inaccurate policy and recover vanilla MPPI-level performance. We theoretically show that re-centering on the prior discards the optimizer's correction at every update, whereas nominal-centered sampling retains it and converges to a solution set by both the task cost and the prior, and that this failure is not removed by a larger rollout budget. Simulations and real-robot experiments demonstrate robust performance under both in-distribution and out-of-distribution tasks.
Road traffic noise remains a major environmental challenge, yet most speed management strategies are static and do not respond to short-term variations in traffic noise emissions. Although variable speed limit (VSL) systems are widely deployed for safety and congestion mitigation, traffic noise is rarely treated as an explicit operational control objective. This paper proposes a noise-aware VSL framework that integrates aggregated traffic-state estimation with a simplified CNOSSOS-EU-based emission indicator. A stage-based controller with time-varying reference thresholds dynamically adjusts discrete speed-limit levels in response to estimated emission conditions. The framework is evaluated using microscopic traffic simulation calibrated with empirical motorway data and replicated across multiple stochastic realisations. Over a 24-hour evaluation period, the adaptive strategy reduces the receiver-based equivalent sound level by 2.9 dB(A) relative to unrestricted traffic conditions, while maintaining an average vehicle speed approximately 11.3 km/h higher than a permanently imposed low-speed regime. Period-wise analysis shows that speed reductions are activated selectively when emission levels approach calibrated targets, rather than enforcing a constant intermediate limit. Traffic stability indicators reveal moderate increases in speed variability compared with unrestricted operation, but substantially lower braking intensity than under uniform low-speed enforcement. These results demonstrate the feasibility of integrating environmental performance indicators into operational speed control, providing a practical complement to conventional infrastructure-based noise mitigation measures.
Early recognition of lane-change intention is essential for proactive decision-making in autonomous driving and advanced driver assistance systems. This paper proposes a Dual Neural-Calibrated Interacting Multiple Model (DNC-IMM) that improves adaptability to driving context while preserving the probabilistic structure and interpretability of a conventional IMM. The proposed method encodes driving-context information, including target-vehicle motion, gaps to surrounding vehicles, and relative velocities, with a neural network that calibrates both the transition-probability matrix and measurement likelihoods. The final intention is determined from the calibrated IMM mode posterior rather than from a separate direct classifier. Experiments on the highD dataset demonstrate that the proposed method reliably recognizes lane-change intentions before lane crossing and provides particularly strong performance at the earlier 2-3 s prediction horizons.
Intelligent vehicles increasingly support adaptive applications beyond driving themselves, ranging from context-aware ADAS and automated driving to in-cabin monitoring and fleet management, all under tight requirements on accuracy, latency, cost, and reliability. Meeting these requirements is challenging because vehicles operate in complex, uncertain, and rapidly changing environments while running on resource-constrained computing platforms. This paper argues that context-situational factors that give meaning to sensor signals and constrain decisions-should be treated as a first-class principle for next-generation vehicle systems, and operationalized as a unified, shared state for learning, risk assessment, and closed-loop control across the software stack. We systematically review state-of-the- art (SOTA) context-aware methods spanning (i) environment understanding, (ii) planning and control, (iii) safety and security, and (iv) connected vehicles. Based on a trend analysis of context-aware design, we identify four key technical challenges in building a general contextual engine for future intelligent vehicles: multi-modal context fusion, temporal context modeling, handling rare events, and collaborative context sharing. We hope this survey will motivate the development of robust and efficient context-aware vehicle applications.
This letter presents a compact two-degree-of-freedom (DoF) robotic finger with a flexion-selective passive continuously variable transmission (CVT) to achieve a wide force-speed operating range. Inspired by the functional differentiation of the human metacarpophalangeal (MCP) joint, the proposed mechanism realizes DoF-specific transmission differentiation by selectively assigning passive CVT to the flexion-extension DoF while preserving direct transmission for abduction-adduction. For a wide force-speed operating range, a force-responsive passive CVT is embedded in the flexion pathway, while direct transmission is preserved for the abduction-adduction pathway. To selectively realize transmission adaptation within a multi-DoF MCP mechanism, an output-side passive CVT employing a moving-pulley-inspired wire-routing structure is introduced. The resultant force generated by the wire tensions acting on the pulley that passively increases the flexion moment arm and transmission ratio according to the applied load without additional actuators, sensors, or control. Experimental results demonstrate a maximum output-force amplification of 4.19-fold and a mean amplification of 3.6-fold across the tested flexion angles ranging from 15 degrees to 75 degrees through moment-arm adaptation, thereby substantially expanding the achievable force-speed operating range. Furthermore, dexterous ball-rolling experiments verify that passive transmission adaptation can be achieved while preserving abduction-adduction functionality. These results demonstrate a scalable transmission design strategy for compact multi-DoF robotic hands.
Conventional soft robot actuators excel in compliance, but their uncontrolled deformations compromise accuracy and hinder scaling to multi-degree-of-freedom (DoF) systems. We introduce a MONOlithic ORIGAMI-inspired soft folding actuator design (MONORIGAMI) that establishes a design strategy based on spatially programmed stiffness anisotropy to preserve material compliance along desired folding directions while selectively restricting deformation in unwanted directions. The actuator leverages stiffness tiers based on material thickness, patterned in an origami-inspired geometry with facets and creases, converting unconstrained soft deformation into accurate, repeatable, and composable folding motions without additional reinforcements. The design is fully 3D-printable through a single-material, single-print process that requires no assembly. Each actuator serves as a scalable motion primitive, and linking and orienting multiple actuators mechanically programs multi-DoF trajectories. Using the same fundamental module, we demonstrate three 3D-printed soft multi-DoF robotic systems spanning distinct application domains: (1) a compact 4-DoF wearable haptic device for high-fidelity cutaneous feedback in virtual reality (VR), (2) a 3-DoF joystick for kinesthetic feedback in teleoperation, and (3) a modular robotic gripper capable of underwater operation with geometry-encoded grasp trajectories. These systems demonstrate the module's capabilities for compact multi-axis integration, controlled physical interaction, and geometry-programmed operation across different environments. Together, these results show that MONORIGAMI provides a general, composable, accessible, reliable, and scalable platform for high-precision soft multi-DoF robotics, addressing long-standing limitations in both soft actuator design and fabrication.
Accurate tilt angle estimation is important in many engineering applications, such as robotics, motion tracking, and embedded control systems. However, measurements from low-cost inertial sensors are often degraded by noise and drift. This paper presents a single-axis tilt angle estimation system based on the MPU6050 inertial measurement unit, implemented on an RP2040 microcontroller platform, with sensor fusion achieved through a Kalman filter. The accelerometer provides a direct estimate of tilt angle from gravity but is sensitive to noise and short-term fluctuations. The gyroscope provides smooth angular rate measurements, but integration over time introduces drift. To overcome these limitations, a Kalman filter is used to combine measurements from both sensors, leveraging the long-term stability of the accelerometer and the short-term smoothness of the gyroscope. Both simulation and hardware experiments are performed. In simulation, sensor noise and drift are modeled to evaluate the filter performance under control conditions. In the hardware implementation, real-time MPU6050 data is acquired and processed by the RP2040 platform, and the estimated tilt angle is compared with accelerometer-only and gyroscope-only outputs. The results show that the proposed method effectively reduces noise measurements and suppresses long-term drift while preserving good dynamic response. Overall, the system provides more stable and accurate tilt estimation than either sensor alone, demonstrating a practical and accessible approach for Kalman filter based sensor fusion in embedded application.
There is tremendous value in humanoid robots taking on physically demanding, hazardous, and repetitive work in spaces built for humans. However, a useful robot for these spaces must coordinate locomotion, whole-body motion, perception, contact, and operator supervision. We present a robot-local, runtime-editable behavior authoring and runtime system that addresses these challenges. We argue that behavior architecture can be a primary enabler of capability, speed, and reliability, and that runtime editability enables fast behavior creation, adaptation, extension, and combination. Our behavior architecture combines object-centric Affordance Templates, a tree structure that provides organization and logic, and runtime-editable perception through a behavior scene and primitive scene actions. Our operator interface remains continuously synchronized to the robot for runtime authoring, monitoring, and repair. Action primitives execute through a whole-body controller that supports concurrent body motions and walking. Demonstrations of our system cover six task variants on Unitree H1-2 and Alex. We execute a push door traversal in 34 seconds and sort six balls by color in 45 seconds under human disturbance. Timed authoring sessions show scratch creation of new loco-manipulation behaviors and adaptation of existing ones in hours. Comparison against the literature finds our approach to be competitive with recent learned systems.
We present ADAPT, an end-to-end framework for interactive, text-conditioned humanoid whole-body control. Unlike dominant text-to-motion pipelines that generate kinematic motions for a separate tracker, ADAPT solves language control with an end-to-end closed-loop control framework, where the robot must continuously respond to changing commands while maintaining balance, natural motion, and smooth transitions. ADAPT learns a diffusion-based action prior from text-labeled humanoid state-action trajectories, enabling diverse motion skills to be directly executed from language commands. To improve long-horizon robustness and smooth prompt switching, we train a lightweight residual reinforcement learning policy on top of the frozen diffusion controller. We further show that the same diffusion policy can be reused as a steerable text-conditioned motion prior for downstream task adaptation. Experiments demonstrate robust language-grounded skill execution, smooth interactive transitions, and style-preserving downstream control.