Multimodal large language models (MLLMs) have achieved impressive success in question-answering tasks, yet their capabilities for spatial understanding are less explored. This work investigates a critical question: do existing MLLMs possess 3D spatial perception and understanding abilities? Concretely, we make the following contributions in this paper: (i) we introduce VGBench, a benchmark specifically designed to assess MLLMs for visual geometry perception, e.g., camera pose and motion estimation; (ii) we propose SpatialScore, the most comprehensive and diverse multimodal spatial understanding benchmark to date, integrating VGBench with relevant data from the other 11 existing datasets. This benchmark comprises 28K samples across various spatial understanding tasks, modalities, and QA formats, along with a carefully curated challenging subset, SpatialScore-Hard; (iii) we develop SpatialAgent, a novel multi-agent system incorporating 9 specialized tools for spatial understanding, supporting both Plan-Execute and ReAct reasoning paradigms; (iv) we conduct extensive evaluations to reveal persistent challenges in spatial reasoning while demonstrating the effectiveness of SpatialAgent. We believe SpatialScore will offer valuable insights and serve as a rigorous benchmark for the next evolution of MLLMs.
Long-term planning, as in reinforcement learning (RL), involves finding strategies: actions that collectively work toward a goal rather than individually optimizing their immediate outcomes. As part of a strategy, some actions are taken at the expense of short-term benefit to enable future actions with even greater returns. These actions are only advantageous if followed up by the actions they facilitate, consequently, they would not have been taken if those follow-ups were not available. In this paper, we quantify such dependencies between planned actions with strategic link scores: the drop in the likelihood of one decision under the constraint that a follow-up decision is no longer available. We demonstrate the utility of strategic link scores through three practical applications: (i) explaining black-box RL agents by identifying strategically linked pairs among decisions they make, (ii) improving the worst-case performance of decision support systems by distinguishing whether recommended actions can be adopted as standalone improvements or whether they are strategically linked hence requiring a commitment to a broader strategy to be effective, and (iii) characterizing the planning processes of non-RL agents purely through interventions aimed at measuring strategic link scores - as an example, we consider a realistic traffic simulator and analyze through road closures the effective planning horizon of the emergent routing behavior of many drivers.
While foundation models (FMs), such as diffusion models and large vision-language models (LVLMs), have been widely applied in educational contexts, their ability to generate pedagogically effective visual explanations remains limited. Most existing approaches focus primarily on textual reasoning, overlooking the critical role of structured and interpretable visualizations in supporting conceptual understanding. To better assess the visual reasoning capabilities of FMs in educational settings, we introduce EduVisBench, a multi-domain, multi-level benchmark. EduVisBench features diverse STEM problem sets requiring visually grounded solutions, along with a fine-grained evaluation rubric informed by pedagogical theory. Our empirical analysis reveals that existing models frequently struggle with the inherent challenge of decomposing complex reasoning and translating it into visual representations aligned with human cognitive processes. To address these limitations, we propose EduVisAgent, a multi-agent collaborative framework that coordinates specialized agents for instructional planning, reasoning decomposition, metacognitive prompting, and visualization design. Experimental results show that EduVisAgent substantially outperforms all baselines, achieving a 40.2% improvement and delivering more educationally aligned visualizations. EduVisBench and EduVisAgent are available at https://github.com/aiming-lab/EduVisBench and https://github.com/aiming-lab/EduVisAgent.
The integration of Vision-Language Models (VLMs) into autonomous driving systems has shown promise in addressing key challenges such as learning complexity, interpretability, and common-sense reasoning. However, existing approaches often struggle with efficient integration and realtime decision-making due to computational demands. In this paper, we introduce SOLVE, an innovative framework that synergizes VLMs with end-to-end (E2E) models to enhance autonomous vehicle planning. Our approach emphasizes knowledge sharing at the feature level through a shared visual encoder, enabling comprehensive interaction between VLM and E2E components. We propose a Trajectory Chain-of-Thought (T-CoT) paradigm, which progressively refines trajectory predictions, reducing uncertainty and improving accuracy. By employing a temporal decoupling strategy, SOLVE achieves efficient cooperation by aligning high-quality VLM outputs with E2E real-time performance. Evaluated on the nuScenes dataset, our method demonstrates significant improvements in trajectory prediction accuracy, paving the way for more robust and reliable autonomous driving systems.
Earth observation foundation models have shown strong generalization across multiple Earth observation tasks, but their robustness under real-world perturbations remains underexplored. To bridge this gap, we introduce REOBench, the first comprehensive benchmark for evaluating the robustness of Earth observation foundation models across six tasks and twelve types of image corruptions, including both appearance-based and geometric perturbations. To ensure realistic and fine-grained evaluation, our benchmark focuses on high-resolution optical remote sensing images, which are widely used in critical applications such as urban planning and disaster response. We conduct a systematic evaluation of a broad range of models trained using masked image modeling, contrastive learning, and vision-language pre-training paradigms. Our results reveal that (1) existing Earth observation foundation models experience significant performance degradation when exposed to input corruptions. (2) The severity of degradation varies across tasks, model architectures, backbone sizes, and types of corruption, with performance drop varying from less than 1% to over 20%. (3) Vision-language models show enhanced robustness, particularly in multimodal tasks. REOBench underscores the vulnerability of current Earth observation foundation models to real-world corruptions and provides actionable insights for developing more robust and reliable models.
Model-based reinforcement learning (MBRL) offers an intuitive way to increase the sample efficiency of model-free RL methods by simultaneously training a world model that learns to predict the future. MBRL methods have progressed by largely prioritising the actor; optimising the world model learning has been neglected meanwhile. Improving the fidelity of the world model and reducing its time to convergence can yield significant downstream benefits, one of which is improving the ensuing performance of any actor it may train. We propose a novel approach that anticipates and actively seeks out high-entropy states using short-horizon latent predictions generated by the world model, offering a principled alternative to traditional curiosity-driven methods that chase once-novel states well after they were stumbled into. While many model predictive control (MPC) based methods offer similar alternatives, they typically lack commitment, synthesising multi step plans after every step. To mitigate this, we present a hierarchical planner that dynamically decides when to replan, planning horizon length, and the weighting between reward and entropy. While our method can theoretically be applied to any model that trains its own actors with solely model generated data, we have applied it to just Dreamer as a proof of concept. Our method finishes the Miniworld procedurally generated mazes 50% faster than base Dreamer at convergence and the policy trained in imagination converges in only 60% of the environment steps that base Dreamer needs.
Remote Sensing Image-Text Retrieval (RSITR) plays a critical role in geographic information interpretation, disaster monitoring, and urban planning by establishing semantic associations between image and textual descriptions. Existing Parameter-Efficient Fine-Tuning (PEFT) methods for Vision-and-Language Pre-training (VLP) models typically adopt symmetric adapter structures for exploring cross-modal correlations. However, the strong discriminative nature of text modality may dominate the optimization process and inhibits image representation learning. The nonnegligible imbalanced cross-modal optimization remains a bottleneck to enhancing the model performance. To address this issue, this study proposes a Representation Discrepancy Bridging (RDB) method for the RSITR task. On the one hand, a Cross-Modal Asymmetric Adapter (CMAA) is designed to enable modality-specific optimization and improve feature alignment. The CMAA comprises a Visual Enhancement Adapter (VEA) and a Text Semantic Adapter (TSA). VEA mines fine-grained image features by Differential Attention (DA) mechanism, while TSA identifies key textual semantics through Hierarchical Attention (HA) mechanism. On the other hand, this study extends the traditional single-task retrieval framework to a dual-task optimization framework and develops a Dual-Task Consistency Loss (DTCL). The DTCL improves cross-modal alignment robustness through an adaptive weighted combination of cross-modal, classification, and exponential moving average consistency constraints. Experiments on RSICD and RSITMD datasets show that the proposed RDB method achieves a 6%-11% improvement in mR metrics compared to state-of-the-art PEFT methods and a 1.15%-2% improvement over the full fine-tuned GeoRSCLIP model.
This paper presents a new approach for 6DoF Direct LiDAR-Inertial Odometry (D-LIO) based on the simultaneous mapping of truncated distance fields on CPU. Such continuous representation (in the vicinity of the points) enables working with raw 3D LiDAR data online, avoiding the need of LiDAR feature selection and tracking, simplifying the odometry pipeline and easily generalizing to many scenarios. The method is based on the proposed Fast Truncated Distance Field (Fast-TDF) method as a convenient tool to represent the environment. Such representation enables i) solving the LiDAR point-cloud registration as a nonlinear optimization process without the need of selecting/tracking LiDAR features in the input data, ii) simultaneously producing an accurate truncated distance field map of the environment, and iii) updating such map at constant time independently of its size. The approach is tested using open datasets, aerial and ground. It is also benchmarked against other state-of-the-art odometry approaches, demonstrating the same or better level of accuracy with the added value of an online-generated TDF representation of the environment, that can be used for other robotics tasks as planning or collision avoidance. The source code is publicly available at https://anonymous.4open.science/r/D-LIO
Existing reasoning datasets saturate and fail to test abstract, multi-step problems, especially pathfinding and complex rule constraint satisfaction. We introduce SPaRC (Spatial Pathfinding Reasoning Challenge), a dataset of 1,000 2D grid pathfinding puzzles to evaluate spatial and symbolic reasoning, requiring step-by-step planning with arithmetic and geometric rules. Humans achieve near-perfect accuracy (98.0%; 94.5% on hard puzzles), while the best reasoning models, such as o4-mini, struggle (15.8%; 1.1% on hard puzzles). Models often generate invalid paths (>50% of puzzles for o4-mini), and reasoning tokens reveal they make errors in navigation and spatial logic. Unlike humans, who take longer on hard puzzles, models fail to scale test-time compute with difficulty. Allowing models to make multiple solution attempts improves accuracy, suggesting potential for better spatial reasoning with improved training and efficient test-time scaling methods. SPaRC can be used as a window into models' spatial reasoning limitations and drive research toward new methods that excel in abstract, multi-step problem-solving.
This paper presents an end-to-end deep reinforcement learning (RL) framework for occlusion-aware robotic manipulation in cluttered plant environments. Our approach enables a robot to interact with a deformable plant to reveal hidden objects of interest, such as fruits, using multimodal observations. We decouple the kinematic planning problem from robot control to simplify zero-shot sim2real transfer for the trained policy. Our results demonstrate that the trained policy, deployed using our framework, achieves up to 86.7% success in real-world trials across diverse initial conditions. Our findings pave the way toward autonomous, perception-driven agricultural robots that intelligently interact with complex foliage plants to "find the fruit" in challenging occluded scenarios, without the need for explicitly designed geometric and dynamic models of every plant scenario.
Maintaining robust 3D perception under dynamic and unpredictable test-time conditions remains a critical challenge for autonomous driving systems. Existing test-time adaptation (TTA) methods often fail in high-variance tasks like 3D object detection due to unstable optimization and sharp minima. While recent model merging strategies based on linear mode connectivity (LMC) offer improved stability by interpolating between fine-tuned checkpoints, they are computationally expensive, requiring repeated checkpoint access and multiple forward passes. In this paper, we introduce CodeMerge, a lightweight and scalable model merging framework that bypasses these limitations by operating in a compact latent space. Instead of loading full models, CodeMerge represents each checkpoint with a low-dimensional fingerprint derived from the source model's penultimate features and constructs a key-value codebook. We compute merging coefficients using ridge leverage scores on these fingerprints, enabling efficient model composition without compromising adaptation quality. Our method achieves strong performance across challenging benchmarks, improving end-to-end 3D detection 14.9% NDS on nuScenes-C and LiDAR-based detection by over 7.6% mAP on nuScenes-to-KITTI, while benefiting downstream tasks such as online mapping, motion prediction and planning even without training. Code and pretrained models are released in the supplementary material.
The automatic control of mobile devices is essential for efficiently performing complex tasks that involve multiple sequential steps. However, these tasks pose significant challenges due to the limited environmental information available at each step, primarily through visual observations. As a result, current approaches, which typically rely on reactive policies, focus solely on immediate observations and often lead to suboptimal decision-making. To address this problem, we propose \textbf{Foresighted Planning with World Model-Driven Code Execution (FPWC)},a framework that prioritizes natural language understanding and structured reasoning to enhance the agent's global understanding of the environment by developing a task-oriented, refinable \emph{world model} at the outset of the task. Foresighted actions are subsequently generated through iterative planning within this world model, executed in the form of executable code. Extensive experiments conducted in simulated environments and on real mobile devices demonstrate that our method outperforms previous approaches, particularly achieving a 44.4\% relative improvement in task success rate compared to the state-of-the-art in the simulated environment. Code and demo are provided in the supplementary material.
Reinforcement Learning (RL) can mitigate the causal confusion and distribution shift inherent to imitation learning (IL). However, applying RL to end-to-end autonomous driving (E2E-AD) remains an open problem for its training difficulty, and IL is still the mainstream paradigm in both academia and industry. Recently Model-based Reinforcement Learning (MBRL) have demonstrated promising results in neural planning; however, these methods typically require privileged information as input rather than raw sensor data. We fill this gap by designing Raw2Drive, a dual-stream MBRL approach. Initially, we efficiently train an auxiliary privileged world model paired with a neural planner that uses privileged information as input. Subsequently, we introduce a raw sensor world model trained via our proposed Guidance Mechanism, which ensures consistency between the raw sensor world model and the privileged world model during rollouts. Finally, the raw sensor world model combines the prior knowledge embedded in the heads of the privileged world model to effectively guide the training of the raw sensor policy. Raw2Drive is so far the only RL based end-to-end method on CARLA Leaderboard 2.0, and Bench2Drive and it achieves state-of-the-art performance.
Inland waterbody detection (IWD) is critical for water resources management and agricultural planning. However, the development of high-fidelity IWD mapping technology remains unresolved. We aim to propose a practical solution based on the easily accessible data, i.e., the delay-Doppler map (DDM) provided by NASA's Cyclone Global Navigation Satellite System (CYGNSS), which facilitates effective estimation of physical parameters on the Earth's surface with high temporal resolution and wide spatial coverage. Specifically, as quantum deep network (QUEEN) has revealed its strong proficiency in addressing classification-like tasks, we encode the DDM using a customized transformer, followed by feeding the transformer-encoded DDM (tDDM) into a highly entangled QUEEN to distinguish whether the tDDM corresponds to a hydrological region. In recent literature, QUEEN has achieved outstanding performances in numerous challenging remote sensing tasks (e.g., hyperspectral restoration, change detection, and mixed noise removal, etc.), and its high effectiveness stems from the fundamentally different way it adopts to extract features (the so-called quantum unitary-computing features). The meticulously designed IWD-QUEEN retrieves high-precision river textures, such as those in Amazon River Basin in South America, demonstrating its superiority over traditional classification methods and existing global hydrography maps. IWD-QUEEN, together with its parallel quantum multihead scheme, works in a near-real-time manner (i.e., millisecond-level computing per DDM). To broaden accessibility for users of traditional computers, we also provide the non-quantum counterpart of our method, called IWD-Transformer, thereby increasing the impact of this work.
Multi-view Synthetic Aperture Radar (SAR) imaging can effectively enhance the performance of tasks such as automatic target recognition and image information fusion. Unmanned aerial vehicles (UAVs) have the advantages of flexible deployment and cost reduction. A swarm of UAVs equipped with synthetic aperture radar imaging equipment is well suited to meet the functional requirements of multi-view synthetic aperture radar imaging missions. However, to provide optimal paths for SAR-UAVs from the base station to cover target viewpoints in the mission area is of NP-hard computational complexity. In this work, the coverage path planning problem for multi-view SAR-UAV observation systems is studied. First, the coordinate of observation viewpoints is calculated based on the location of targets and base station under a brief geometric model. Then, the exact problem formulation is modeled in order to fully describe the solution space and search for optimal paths that provide maximum coverage rate for SAR-UAVs. Finally, an Adaptive Density Peak Clustering (ADPC) method is proposed to overcome the additional energy consumption due to the viewpoints being far away from the base station. The Particle Swarm Optimization (PSO) algorithm is introduced for optimal path generation. Experimental results demonstrate the effectiveness and computational efficiency of the proposed approach.
Reinforcement learning (RL)-based autonomous driving policy learning faces critical limitations such as low sample efficiency and poor generalization; its reliance on online interactions and trial-and-error learning is especially unacceptable in safety-critical scenarios. Existing methods including safe RL often fail to capture the true semantic meaning of "safety" in complex driving contexts, leading to either overly conservative driving behavior or constraint violations. To address these challenges, we propose VL-SAFE, a world model-based safe RL framework with Vision-Language model (VLM)-as-safety-guidance paradigm, designed for offline safe policy learning. Specifically, we construct offline datasets containing data collected by expert agents and labeled with safety scores derived from VLMs. A world model is trained to generate imagined rollouts together with safety estimations, allowing the agent to perform safe planning without interacting with the real environment. Based on these imagined trajectories and safety evaluations, actor-critic learning is conducted under VLM-based safety guidance to optimize the driving policy more safely and efficiently. Extensive evaluations demonstrate that VL-SAFE achieves superior sample efficiency, generalization, safety, and overall performance compared to existing baselines. To the best of our knowledge, this is the first work that introduces a VLM-guided world model-based approach for safe autonomous driving. The demo video and code can be accessed at: https://ys-qu.github.io/vlsafe-website/
Automatic detection of speech dysfluency aids speech-language pathologists in efficient transcription of disordered speech, enhancing diagnostics and treatment planning. Traditional methods, often limited to classification, provide insufficient clinical insight, and text-independent models misclassify dysfluency, especially in context-dependent cases. This work introduces Dysfluent-WFST, a zero-shot decoder that simultaneously transcribes phonemes and detects dysfluency. Unlike previous models, Dysfluent-WFST operates with upstream encoders like WavLM and requires no additional training. It achieves state-of-the-art performance in both phonetic error rate and dysfluency detection on simulated and real speech data. Our approach is lightweight, interpretable, and effective, demonstrating that explicit modeling of pronunciation behavior in decoding, rather than complex architectures, is key to improving dysfluency processing systems.
Manipulating elasto-plastic objects remains a significant challenge due to severe self-occlusion, difficulties of representation, and complicated dynamics. This work proposes a novel framework for elasto-plastic object manipulation with a quasi-static assumption for motions, leveraging 3D occupancy to represent such objects, a learned dynamics model trained with 3D occupancy, and a learning-based predictive control algorithm to address these challenges effectively. We build a novel data collection platform to collect full spatial information and propose a pipeline for generating a 3D occupancy dataset. To infer the 3D occupancy during manipulation, an occupancy prediction network is trained with multiple RGB images supervised by the generated dataset. We design a deep neural network empowered by a 3D convolution neural network (CNN) and a graph neural network (GNN) to predict the complex deformation with the inferred 3D occupancy results. A learning-based predictive control algorithm is introduced to plan the robot actions, incorporating a novel shape-based action initialization module specifically designed to improve the planner efficiency. The proposed framework in this paper can successfully shape the elasto-plastic objects into a given goal shape and has been verified in various experiments both in simulation and the real world.
Validating scientific hypotheses is a central challenge in biomedical research, and remains difficult for artificial intelligence (AI) agents due to the complexity of real-world data analysis and evidence interpretation. In this work, we present BioDSA-1K, a benchmark designed to evaluate AI agents on realistic, data-driven biomedical hypothesis validation tasks. BioDSA-1K consists of 1,029 hypothesis-centric tasks paired with 1,177 analysis plans, curated from over 300 published biomedical studies to reflect the structure and reasoning found in authentic research workflows. Each task includes a structured hypothesis derived from the original study's conclusions, expressed in the affirmative to reflect the language of scientific reporting, and one or more pieces of supporting evidence grounded in empirical data tables. While these hypotheses mirror published claims, they remain testable using standard statistical or machine learning methods. The benchmark enables evaluation along four axes: (1) hypothesis decision accuracy, (2) alignment between evidence and conclusion, (3) correctness of the reasoning process, and (4) executability of the AI-generated analysis code. Importantly, BioDSA-1K includes non-verifiable hypotheses: cases where the available data are insufficient to support or refute a claim, reflecting a common yet underexplored scenario in real-world science. We propose BioDSA-1K as a foundation for building and evaluating generalizable, trustworthy AI agents for biomedical discovery.
Discovering regularities from spatiotemporal systems can benefit various scientific and social planning. Current spatiotemporal learners usually train an independent model from a specific source data that leads to limited transferability among sources, where even correlated tasks requires new design and training. The key towards increasing cross-domain knowledge is to enable collective intelligence and model evolution. In this paper, inspired by neuroscience theories, we theoretically derive the increased information boundary via learning cross-domain collective intelligence and propose a Synaptic EVOlutional spatiotemporal network, SynEVO, where SynEVO breaks the model independence and enables cross-domain knowledge to be shared and aggregated. Specifically, we first re-order the sample groups to imitate the human curriculum learning, and devise two complementary learners, elastic common container and task-independent extractor to allow model growth and task-wise commonality and personality disentanglement. Then an adaptive dynamic coupler with a new difference metric determines whether the new sample group should be incorporated into common container to achieve model evolution under various domains. Experiments show that SynEVO improves the generalization capacity by at most 42% under cross-domain scenarios and SynEVO provides a paradigm of NeuroAI for knowledge transfer and adaptation.
Extremum Seeking (ES) is an effective real-time optimization method for PDE systems in cascade with nonlinear quadratic maps. To address PDEs in the feedback loop, a boundary control law and a re-design of the additive probing signal are mandatory. The latter, commonly called "trajectory generation" or "motion planning," involves designing perturbation signals that anticipate their propagation through PDEs. Specifically, this requires solving motion planning problems for systems governed by parabolic and hyperbolic PDEs. Physics-Informed Neural Networks (PINN) is a powerful tool for solving PDEs by embedding physical laws as constraints in the neural network's loss function, enabling efficient solutions for high-dimensional, nonlinear, and complex problems. This paper proposes a novel construction integrating PINN and ES, automating the motion planning process for specific PDE systems and eliminating the need for case-by-case analytical derivations. The proposed strategy efficiently extracts perturbation signals, optimizing the PDE system.
While autonomous driving (AD) stacks struggle with decision making under partial observability and real-world complexity, human drivers are capable of commonsense reasoning to make near-optimal decisions with limited information. Recent work has attempted to leverage finetuned Vision-Language Models (VLMs) for trajectory planning at inference time to emulate human behavior. Despite their success in benchmark evaluations, these methods are often impractical to deploy (a 70B parameter VLM inference at merely 8 tokens per second requires more than 160G of memory), and their monolithic network structure prohibits safety decomposition. To bridge this gap, we propose VLM-Embedded Reasoning for autonomous Driving (VERDI), a training-time framework that distills the reasoning process and commonsense knowledge of VLMs into the AD stack. VERDI augments modular differentiable end-to-end (e2e) AD models by aligning intermediate module outputs at the perception, prediction, and planning stages with text features explaining the driving reasoning process produced by VLMs. By encouraging alignment in latent space, \textsc{VERDI} enables the modular AD stack to internalize structured reasoning, without incurring the inference-time costs of large VLMs. We demonstrate the effectiveness of our method on the NuScenes dataset and find that VERDI outperforms existing e2e methods that do not embed reasoning by 10% in $\ell_{2}$ distance, while maintaining high inference speed.
Planning the transition to quantum-safe cryptosystems requires understanding the cost of quantum attacks on vulnerable cryptosystems. In Gidney+Eker{\aa} 2019, I co-published an estimate stating that 2048 bit RSA integers could be factored in eight hours by a quantum computer with 20 million noisy qubits. In this paper, I substantially reduce the number of qubits required. I estimate that a 2048 bit RSA integer could be factored in less than a week by a quantum computer with less than a million noisy qubits. I make the same assumptions as in 2019: a square grid of qubits with nearest neighbor connections, a uniform gate error rate of $0.1\%$, a surface code cycle time of 1 microsecond, and a control system reaction time of $10$ microseconds. The qubit count reduction comes mainly from using approximate residue arithmetic (Chevignard+Fouque+Schrottenloher 2024), from storing idle logical qubits with yoked surface codes (Gidney+Newman+Brooks+Jones 2023), and from allocating less space to magic state distillation by using magic state cultivation (Gidney+Shutty+Jones 2024). The longer runtime is mainly due to performing more Toffoli gates and using fewer magic state factories compared to Gidney+Eker{\aa} 2019. That said, I reduce the Toffoli count by over 100x compared to Chevignard+Fouque+Schrottenloher 2024.
In this work, we contribute the first approach to solve infinite-horizon discounted general-utility Markov decision processes (GUMDPs) in the single-trial regime, i.e., when the agent's performance is evaluated based on a single trajectory. First, we provide some fundamental results regarding policy optimization in the single-trial regime, investigating which class of policies suffices for optimality, casting our problem as a particular MDP that is equivalent to our original problem, as well as studying the computational hardness of policy optimization in the single-trial regime. Second, we show how we can leverage online planning techniques, in particular a Monte-Carlo tree search algorithm, to solve GUMDPs in the single-trial regime. Third, we provide experimental results showcasing the superior performance of our approach in comparison to relevant baselines.
Continuous time systems are often modeled using discrete time dynamics but this requires a small simulation step to maintain accuracy. In turn, this requires a large planning horizon which leads to computationally demanding planning problems and reduced performance. Previous work in model free reinforcement learning has partially addressed this issue using action repeats where a policy is learned to determine a discrete action duration. Instead we propose to control the continuous decision timescale directly by using temporally-extended actions and letting the planner treat the duration of the action as an additional optimization variable along with the standard action variables. This additional structure has multiple advantages. It speeds up simulation time of trajectories and, importantly, it allows for deep horizon search in terms of primitive actions while using a shallow search depth in the planner. In addition, in the model based reinforcement learning (MBRL) setting, it reduces compounding errors from model learning and improves training time for models. We show that this idea is effective and that the range for action durations can be automatically selected using a multi-armed bandit formulation and integrated into the MBRL framework. An extensive experimental evaluation both in planning and in MBRL, shows that our approach yields faster planning, better solutions, and that it enables solutions to problems that are not solved in the standard formulation.
This paper presents a distributionally robust planning method for hydrogen-electrical microgrids over islands, where the cross-island energy exchange is supported by a maritime hydrogen transport network. This planning problem is complicated due to heterogeneous off-shore wind-driven uncertainties (i.e., renewable power, transport availability, demand fluctuations, and grid faulting), a subset of which exhibit endogenous uncertainty, as they can be affected by proactive measures (e.g., grid hardening) or infrastructure investment. To capture these features, a two-stage distributionally robust optimization (DRO) model is developed considering decision-dependent uncertainty (DDU), which encompasses variation of the underlying distributional ambiguity due to the change of the first stage decisions. Notably, the complete recourse property is missing, which is often neglected in existing DRO studies. Nevertheless, different from the case for land-based microgrids, this issue is critical and fundamental for sea island systems due to their particular physical and logistical requirements. To address these issues, we develop a C&CG algorithm that is customized with strong cutting planes to handle DRO with a varying DDU ambiguity set and feasibility requirements. Numerical results demonstrate the cost-effectiveness and resilience of the proposed planning framework, along with the nontrivial improvements of the algorithm in both solution accuracy and computational efficiency.
Unmanned Aerial Vehicles (UAVs) are evolving into language-interactive platforms, enabling more intuitive forms of human-drone interaction. While prior works have primarily focused on high-level planning and long-horizon navigation, we shift attention to language-guided fine-grained trajectory control, where UAVs execute short-range, reactive flight behaviors in response to language instructions. We formalize this problem as the Flying-on-a-Word (Flow) task and introduce UAV imitation learning as an effective approach. In this framework, UAVs learn fine-grained control policies by mimicking expert pilot trajectories paired with atomic language instructions. To support this paradigm, we present UAV-Flow, the first real-world benchmark for language-conditioned, fine-grained UAV control. It includes a task formulation, a large-scale dataset collected in diverse environments, a deployable control framework, and a simulation suite for systematic evaluation. Our design enables UAVs to closely imitate the precise, expert-level flight trajectories of human pilots and supports direct deployment without sim-to-real gap. We conduct extensive experiments on UAV-Flow, benchmarking VLN and VLA paradigms. Results show that VLA models are superior to VLN baselines and highlight the critical role of spatial grounding in the fine-grained Flow setting.
Autonomous underwater vehicles (AUVs) are increasingly used in marine research, military applications, and undersea exploration. However, their operational range is significantly affected by battery performance. In this paper, a framework for a wireless energy sharing system among AUVs is proposed, enabling rapid energy replenishment. Path planning plays a crucial role in the energy-sharing process and autonomous navigation, as it must generate feasible trajectories toward designated goals. This article focuses on efficient obstacle avoidance in complex underwater environments, including irregularly shaped obstacles and narrow passages. The proposed method combines Rapidly-exploring Random Trees Star (RRT*) with Particle Swarm Optimization (PSO) to improve path planning efficiency. Comparative analysis of the two algorithms is presented through simulation results in both random and irregular obstacle environments. Index Terms: Wireless charging, autonomous underwater vehicles (AUVs), path planning, irregular obstacles, narrow passages, RRT*, particle swarm optimization (PSO).
Swarm robotic trajectory planning faces challenges in computational efficiency, scalability, and safety, particularly in complex, obstacle-dense environments. To address these issues, we propose SwarmDiff, a hierarchical and scalable generative framework for swarm robots. We model the swarm's macroscopic state using Probability Density Functions (PDFs) and leverage conditional diffusion models to generate risk-aware macroscopic trajectory distributions, which then guide the generation of individual robot trajectories at the microscopic level. To ensure a balance between the swarm's optimal transportation and risk awareness, we integrate Wasserstein metrics and Conditional Value at Risk (CVaR). Additionally, we introduce a Diffusion Transformer (DiT) to improve sampling efficiency and generation quality by capturing long-range dependencies. Extensive simulations and real-world experiments demonstrate that SwarmDiff outperforms existing methods in computational efficiency, trajectory validity, and scalability, making it a reliable solution for swarm robotic trajectory planning.
Vision-Language Models (VLMs) acquire real-world knowledge and general reasoning ability through Internet-scale image-text corpora. They can augment robotic systems with scene understanding and task planning, and assist visuomotor policies that are trained on robot trajectory data. We explore the reverse paradigm - using rich, real, multi-modal robot trajectory data to enhance and evaluate VLMs. In this paper, we present Robo2VLM, a Visual Question Answering (VQA) dataset generation framework for VLMs. Given a human tele-operated robot trajectory, Robo2VLM derives ground-truth from non-visual and non-descriptive sensory modalities, such as end-effector pose, gripper aperture, and force sensing. Based on these modalities, it segments the robot trajectory into a sequence of manipulation phases. At each phase, Robo2VLM uses scene and interaction understanding to identify 3D properties of the robot, task goal, and the target object. The properties are used to generate representative VQA queries - images with textural multiple-choice questions - based on spatial, goal-conditioned, and interaction reasoning question templates. We curate Robo2VLM-1, a large-scale in-the-wild dataset with 684,710 questions covering 463 distinct scenes and 3,396 robotic manipulation tasks from 176k real robot trajectories. Results suggest that Robo2VLM-1 can benchmark and improve VLM capabilities in spatial and interaction reasoning.