Planning

Robotic path planning problems are often NP-hard, and practical solutions typically rely on approximation algorithms with provable performance guarantees for general cases. While designing such algorithms is challenging, formally proving their approximation optimality is even more demanding, which requires domain-specific geometric insights and multi-step mathematical reasoning over complex operational constraints. Recent Large Language Models (LLMs) have demonstrated strong performance on mathematical reasoning benchmarks, yet their ability to assist with research-level optimality proofs in robotic path planning remains under-explored. In this work, we introduce the first benchmark for evaluating LLMs on approximation-ratio proofs of robotic path planning algorithms. The benchmark consists of 34 research-grade proof tasks spanning diverse planning problem types and complexity levels, each requiring structured reasoning over algorithm descriptions, problem constraints, and theoretical guarantees. Our evaluation of state-of-the-art proprietary and open-source LLMs reveals that even the strongest models struggle to produce fully valid proofs without external domain knowledge. However, providing LLMs with task-specific in-context lemmas substantially improves reasoning quality, a factor that is more effective than generic chain-of-thought prompting or supplying the ground-truth approximation ratio as posterior knowledge. We further provide fine-grained error analysis to characterize common logical failures and hallucinations, and demonstrate how each error type can be mitigated through targeted context augmentation.

Communication is a core enabler for multi-robot systems (MRS), providing the mechanism through which robots exchange state information, coordinate actions, and satisfy safety constraints. While many MRS autonomy algorithms assume reliable and timely message delivery, realistic wireless channels introduce delay, erasures, and ordering stalls that can degrade performance and compromise safety-critical decisions of the robot task. In this paper, we investigate how transport-layer reliability mechanisms that mitigate communication losses and delays shape the autonomy-communication loop. We show that conventional non-coded retransmission-based protocols introduce long delays that are misaligned with the timeliness requirements of MRS applications, and may render the received data irrelevant. As an alternative, we advocate for adaptive and causal network coding, which proactively injects coded redundancy to achieve the desired delay and throughput, enabling relevant data delivery for the robotic task. Specifically, this method adapts to channel conditions between robots and causally tunes the communication rates via efficient algorithms. We present two case studies: cooperative localization under delayed and lossy inter-robot communication, and a safety-critical overtaking maneuver where timely vehicle-to-vehicle message availability determines whether an ego vehicle can abort to avoid a crash. Our results demonstrate that coding-based communication significantly reduces in-order delivery stalls, keeps cooperative-localization accuracy close to the ideal baseline, and satisfies the overtaking abort deadline in 80% of the simulated runs, compared with 60% for a retransmission-based baseline. The study highlights the need to jointly design autonomy algorithms and communication mechanisms, and positions network coding as a principled tool for dependable MRS operation over wireless networks.

Remote sensing underpins crucial applications such as disaster relief and ecological field surveys, where systems must understand complex scenes and constraints and make reliable decisions. Current remote-sensing benchmarks mainly focus on evaluating perception and reasoning capabilities of multimodal large language models (MLLMs). They fail to assess planning capability, stemming either from the difficulty of curating and validating planning tasks at scale or from evaluation protocols that are inaccurate and inadequate. To address these limitations, we introduce NeSy-Route, a large-scale neuro-symbolic benchmark for constrained route planning in remote sensing. Within this benchmark, we introduce an automated data-generation framework that integrates high-fidelity semantic masks with heuristic search to produce diverse route-planning tasks with provably optimal solutions. This allows NeSy-Route to comprehensively evaluate planning across 10,821 route-planning samples, nearly 10 times larger than the largest prior benchmark. Furthermore, a three-level hierarchical neuro-symbolic evaluation protocol is developed to enable accurate assessment and support fine-grained analysis on perception, reasoning, and planning simultaneously. Our comprehensive evaluation of various state-of-the-art MLLMs demonstrates that existing MLLMs show significant deficiencies in perception and planning capabilities. We hope NeSy-Route can support further research and development of more powerful MLLMs for remote sensing. The dataset and code are available at https://mingyang1010.github.io/NeSy-Route/.

Motion planning under dynamics constraints, i.e, kinodynamic planning, enables safe robot operation by generating dynamically feasible trajectories that the robot can accurately track. For high-DOF robots such as manipulators, sampling-based motion planners are commonly used, especially for complex tasks in cluttered environments. However, enforcing constraints on robot dynamics in such planners requires solving either challenging two-point boundary value problems (BVPs) or propagating robot dynamics, both of which cause computational bottlenecks that drastically increase planning times. Meanwhile, recent efforts have shown that sampling-based motion planners can generate plans in microseconds using parallelization, but are limited to geometric paths. This paper develops FLASK, a fast parallelized sampling-based kinodynamic motion planning framework for a broad class of differentially flat robot systems, including manipulators, ground and aerial vehicles, and more. Differential flatness allows us to transform the motion planning problem from the original state space to a flat output space, where an analytical time-parameterized solution of the BVP problem can be obtained. A trajectory in the flat output space is then converted back to a closed-form dynamically feasible trajectory in the original state space, enabling fast validation via ``single instruction, multiple data" parallelism. Our framework is fast, exact, and compatible with any sampling-based motion planner, while offering theoretical guarantees on probabilistic exhaustivity and asymptotic optimality based on the closed-form BVP solutions. We extensively verify the effectiveness of our approach in both simulated benchmarks and real experiments with cluttered and dynamic environments, requiring mere microseconds to milliseconds of planning time.

Non-prehensile manipulation is essential for handling thin, large, or otherwise ungraspable objects in unstructured settings. Prior planning and search-based methods often rely on ad-hoc manual designs or generate physically unrealizable motions by ignoring critical gripper properties, while training-based approaches are data-intensive and struggle to generalize to novel, out-of-distribution tasks. We propose a library-driven hierarchical planner (LDHP) that makes executability a first-class design goal: a top-tier contact-state planner proposes object-pose paths using MoveObject primitives, and a bottom-tier grasp planner synthesizes feasible grasp sequences with AdjustGrasp primitives; feasibility is certified by collision checks and quasi-static mechanics, and contact-sensitive segments are recovered via a bounded dichotomy refinement. This gripper-aware decomposition decouples object motion from grasp realizability, yields a task-agnostic pipeline that transfers across manipulation tasks and geometric variations without re-design, and exposes clean hooks for optional learned priors. Real-robot studies on zero-mobility lifting and slot insertion demonstrate consistent execution and robustness to shape and environment changes.

Real-world robots often operate in settings where objective priorities depend on the underlying context of operation. When the underlying context is unknown apriori, multiple robots may have to coordinate to gather informative observations to infer the context, since acting based on an incorrect context can lead to misaligned and unsafe behavior. Once the underlying true context is inferred, the robots optimize their task-specific objectives in the preference order induced by the context. We formalize this problem as a Multi-Robot Context-Uncertain Stochastic Shortest Path (MR-CUSSP), which captures context-relevant information at landmark states through joint observations. Our two-stage solution approach is composed of: (1) CIMOP (Coordinated Inference for Multi-Objective Planning) to compute plans that guide robots toward informative landmarks to efficiently infer the true context, and (2) LCBS (Lexicographic Conflict-Based Search) for collision-free multi-robot path planning with lexicographic objective preferences, induced by the context. We evaluate the algorithms using three simulated domains and demonstrate its practical applicability using five mobile robots in the salp domain setup.

Motion planning through narrow passages remains a core challenge: sampling-based planners rarely place samples inside these narrow but critical regions, and even when samples land inside a passage, the straight-line connections between them run close to obstacle boundaries and are frequently rejected by collision checking. Decomposition-based planners resolve both issues by partitioning free space into convex cells -- every passage is captured exactly as a cell boundary, and any path within a cell is collision-free by construction. However, the number of candidate corridors through the cell graph grows combinatorially with environment complexity, creating a bottleneck in corridor selection. We present GNN-DIP, a framework that addresses this by integrating a Graph Neural Network (GNN) with a two-phase Decomposition-Informed Planner (DIP). The GNN predicts portal scores on the cell adjacency graph to bias corridor search toward near-optimal regions while preserving completeness. In 2D, Constrained Delaunay Triangulation (CDT) with the Funnel algorithm yields exact shortest paths within corridors; in 3D, Slab convex decomposition with portal-face sampling provides near-optimal path evaluation. Benchmarks on 2D narrow-passage scenarios, 3D bottleneck environments with up to 246 obstacles, and dynamic 2D settings show that GNN-DIP achieves 99--100% success rates with 2--280 times speedup over sampling-based baselines.

This paper proposes SoftGM, an octopus-inspired distributed control architecture for segmented soft robotic arms that learn to reach targets in contact-rich environments using online obstacle discovery without assuming full obstacle geometry at the start of each episode. SoftGM formulates each arm section as a cooperative agent and represents the arm-environment interaction as a graph. SoftGM uses a two-stage graph attention message passing scheme following a Centralised Training Decentralised Execution (CTDE) paradigm with a centralised critic and decentralised actor. We evaluate SoftGM in a Cosserat-rod simulator (PyElastica) across three tasks that increase the complexity of the environment: obstacle-free, structured obstacles, and a wall-with-hole scenario. Compared with six widely used MARL baselines (IDDPG, IPPO, ISAC, MADDPG, MAPPO, MASAC) under identical information content and training conditions, SoftGM matches strong CTDE methods in simpler settings and achieves the best performance in the wall-with-hole task. Robustness tests with observation noise, single-section actuation failure, and transient disturbances show that SoftGM maintains comparable performance under non-ideal simulated conditions while keeping control effort bounded, suggesting that selective contact-relevant information routing improves resilience in the tested settings.

Intelligent agents must reason over both continuous dynamics and discrete representations to generate effective plans in complex environments. Previous studies have shown that symbolic abstractions can emerge from neural effect predictors trained with a robot's unsupervised exploration. However, these methods rely on deterministic symbolic domains, lack mechanisms to verify the generated symbolic plans, and operate only at the abstract level, often failing to capture the continuous dynamics of the environment. To overcome these limitations, we propose a bilevel neuro-symbolic framework in which learned probabilistic symbolic rules generate candidate plans rapidly at the high level, and learned continuous effect models verify these plans and perform forward search when necessary at the low level. Our experiments on multi-object manipulation tasks demonstrate that the proposed bilevel method outperforms symbolic-only approaches, reliably identifying failing plans through verification, and achieves planning performance statistically comparable to continuous forward search while resolving most problems via efficient symbolic reasoning.

This paper bridges some of the gap between optimal planning and reinforcement learning (RL), both of which share roots in dynamic programming applied to sequential decision making or optimal control. Whereas planning typically favors deterministic models, goal termination, and cost minimization, RL tends to favor stochastic models, infinite-horizon discounting, and reward maximization in addition to learning-related parameters such as the learning rate and greediness factor. A derandomized version of RL is developed, analyzed, and implemented to yield performance comparisons with value iteration and Dijkstra's algorithm using simple planning models. Next, mathematical analysis shows: 1) conditions under which cost minimization and reward maximization are equivalent, 2) conditions for equivalence of single-shot goal termination and infinite-horizon episodic learning, and 3) conditions under which discounting causes goal achievement to fail. The paper then advocates for defining and optimizing truecost, rather than inserting arbitrary parameters to guide operations. Performance studies are then extended to the stochastic case, using planning-oriented criteria and comparing value iteration to RL with learning rates and greediness factors.

Developing autonomous mobile robot systems typically requires either extensive, platform-specific data collection or relies on simplified abstractions, such as unicycle or bicycle models, that fail to capture the complex kinodynamics of diverse platforms, ranging from wheeled to tracked vehicles. This limitation hinders scalability across evolving heterogeneous autonomous robot fleets. To address this challenge, we propose Cross-vehicle kinodynamics Adaptation via mobility Representation (CAR), a novel framework that enables rapid mobility transfer to new vehicles. CAR employs a Transformer encoder with Adaptive Layer Normalization to embed vehicle trajectory transitions and physical configurations into a shared mobility latent space. By identifying and extracting commonality from nearest neighbors within this latent space, our approach enables rapid kinodynamics adaptation to novel platforms with minimal data collection and computational overhead. We evaluate CAR using the Verti-Bench simulator, built on the Chrono multi-physics engine, and validate its performance on four distinct physical configurations of the Verti-4-Wheeler platform. With only one minute of new trajectory data, CAR achieves up to 67.2% reduction in prediction error compared to direct neighbor transfer across diverse unseen vehicle configurations, demonstrating the effectiveness of cross-vehicle mobility knowledge transfer in both simulated and real-world environments.

Large language models have been widely used for robotic task planning, often taking advantage of representations such as Behavior Trees (BTs). Vision-Language Models (VLMs) have extended these works by grounding the generated plans in the observed scene. However, existing methods are either text-only or rely on large proprietary VLMs, while no dataset pairs visual observations and task instructions with executable and ROS2-compatible BTs. We address this gap with a multi-stage teacher pipeline that converts 1,622 Open X-Embodiment episodes into an augmented multimodal BT dataset containing 2,433 examples. We use this dataset to fine-tune compact and open-source VLMs, ranging from 500M to 4B parameters, using parameter-efficient fine-tuning (PEFT). We then evaluate the generated BTs offline in terms of syntactic correctness and by executing them on 15 household tasks in BEHAVIOR-1K. Our best model, Gemma-3 4B, achieves perfect BT validity and an 87% success rate, outperforming Claude Opus 4.8 and approaching GPT-5, while running locally. Finally, our ablation studies show that adding visual observations increases task success from 40% to 87%, while data augmentation increases BT validity from 65% to 100%.

Autonomous robotic systems are increasingly deployed for mapping, monitoring, and inspection in complex and unstructured environments. However, most existing path planning approaches remain domain-specific (i.e., in the air, on land, or at sea), limiting their scalability and cross-platform applicability. This article presents OmniPlanner, a unified planning framework for autonomous exploration and inspection across aerial, ground, and underwater robots. The method integrates volumetric exploration and viewpoint-based inspection, alongside target reach behaviors within a single modular architecture, complemented by a platform abstraction layer that captures morphology-specific sensing, traversability and motion constraints. This enables the same planning strategy to generalize across distinct mobility domains with minimal retuning. The framework is validated through extensive simulation studies and field deployments in underground mines, industrial facilities, forests, submarine bunkers, and structured outdoor environments. Across these diverse scenarios, OmniPlanner demonstrates robust performance, consistent cross-domain generalization, and improved exploration and inspection efficiency compared to representative state-of-the-art baselines. Videos presenting the OmniPlanner framework and demonstrating its field deployments across aerial, ground, and underwater robots are available at https://ntnu-arl.github.io/omniplanner, and the source code is publicly available at https://github.com/ntnu-arl/gbplanner_ros/tree/gbplanner3.

When automating plan generation for a real-world sequential decision problem, the goal is often not to replace the human planner, but to facilitate an iterative reasoning and elicitation process, where the human's role is to guide the AI planner according to their preferences and expertise. In this context, explanations that respond to users' questions are crucial to improve their understanding of potential solutions and increase their trust in the system. To enable natural interaction with such a system, we present a multi-agent Large Language Model (LLM) architecture that is agnostic to the explanation framework and enables user- and context-dependent interactive explanations. We also describe an instantiation of this framework for goal-conflict explanations, which we use to conduct a user study comparing the LLM-powered interaction with a baseline template-based explanation interface.

In this paper, we present an online method for converting an arbitrary geometric path, represented by a sequence of states, and generated by any planner (e.g., sampling-based planners such as RRT or PRM, search-based planners such as ARA*, etc.), into a kinematically feasible, jerk-limited trajectory. The method generates a sequence of quintic/quartic splines that can be discretized at a user-specified control rate and streamed to a low-level robot controller. Our approach enables real-time adaptation to environmental changes and can be re-invoked at any instant to generate a new trajectory from the robot's current state to a desired target state or sequence of states. Under a bounded-obstacle-velocity assumption, the method provides conditional stopping-safety guarantees over a finite time interval in dynamic environments, while allowing bounded geometric deviation from the original path. Kinematic constraints, including jerk limits, are explicitly considered. We validate the approach in a comparative simulation study against a competing method, demonstrating favorable behavior w.r.t. smoothness, computational time, and real-time performance, particularly with frequent target-state changes (up to 1 [kHz]). Real-robot experiments demonstrate applicability in real-world scenarios, including scenarios with a human as an obstacle.

With the recent fast development of generative models, instruction-based image editing has shown great potential in generating high-quality images. However, the quality of editing highly depends on carefully designed instructions, placing the burden of task decomposition and sequencing entirely on the user. To achieve autonomous image editing, we present PhotoAgent, a system that advances image editing through explicit aesthetic planning. Specifically, PhotoAgent formulates autonomous image editing as a long-horizon decision-making problem. It reasons over user aesthetic intent, plans multi-step editing actions via tree search, and iteratively refines results through closed-loop execution with memory and visual feedback, without requiring step-by-step user prompts. To support reliable evaluation in real-world scenarios, we introduce UGC-Edit, an aesthetic evaluation benchmark consisting of 7,000 photos and a learned aesthetic reward model. We also construct a test set containing 1,017 photos to systematically assess autonomous photo editing performance. Extensive experiments demonstrate that PhotoAgent consistently improves both instruction adherence and visual quality compared with baseline methods. The project page is https://mdyao.github.io/PhotoAgent/.

This paper investigates a problem called Multi-Agent Path Finding with Elevators (MAPF-E), which seeks conflict-free paths for multiple agents from their start to goal locations that may locate on different floors, and the agents can use elevators to travel between floors. The existence of elevators complicates the interaction among the agents and introduces new challenges to the planning. On the one hand, elevators can cause many conflicts among the agents due to its relatively long traversal time across floors, especially when many agents need to reach a different floor. On the other hand, the planner has to reason in a larger state space including the states of the elevators, besides the locations of the agents.

Autonomous aerial scanning of target structures is crucial for practical applications, requiring online adaptation to unknown obstacles during flight. Existing methods largely emphasize collision avoidance and efficiency, but overlook occlusion-induced visibility degradation, severely compromising scanning quality. This study proposes FC-Vision, an on-the-fly visibility-aware replanning framework that proactively and safely prevents target occlusions while preserving full target coverage and efficiency of the original plan. Our approach explicitly enforces dense surface-visibility constraints to regularize replanning behavior in real-time via an efficient two-level decomposition: occlusion-free viewpoint repair that maintains coverage with minimal deviation from the nominal scan, followed by segment-wise clean-sensing connection in 5-DoF space. A plug-in integration strategy is also presented to seamlessly interface \textbf{FC-Vision} with existing UAV scanning systems without architectural changes. Comprehensive simulation and real-world evaluations show that \textbf{FC-Vision} consistently improves scanning quality under unexpected occluders, delivering a maximum coverage gain of 55.32% and a 73.17% reduction in the occlusion ratio, while achieving real-time performance with a moderate increase in flight time. The code has been released at https://github.com/FC-Family/FC-Vision.

Multi-Agent Path Finding (MAPF) is a core coordination problem for large robot fleets in automated warehouses and logistics. Existing approaches are typically either open-loop planners, which must compute complete trajectories before execution and therefore may incur substantial planning latency before actions can be taken, or closed-loop heuristics without reliable performance guarantees, limiting their use in safety-critical deployments. This paper presents Anytime Closed-Loop Conflict-Based Search (ACCBS), a closed-loop algorithm built on a finite-horizon variant of Conflict-Based Search (CBS) with a horizon-changing mechanism inspired by iterative horizon-deepening in Model Predictive Control (MPC). ACCBS dynamically adjusts the planning horizon based on the available computational budget, and reuses a single constraint tree to enable seamless transitions between horizons. As a result, it produces high-quality feasible solutions quickly while being asymptotically optimal as the budget increases, exhibiting anytime behavior. Extensive case studies demonstrate that ACCBS achieves a favorable balance between computational efficiency, solution quality, and execution flexibility, while naturally accommodating online disturbances through its closed-loop formulation.

Visual navigation models often struggle in real-world dynamic environments due to limited robustness to the sim-to-real gap and the difficulty of training policies tailored to target deployment environments (e.g., households, restaurants, and factories). Although real-to-sim navigation simulation using 3D Gaussian Splatting (GS) can mitigate these challenges, prior GS-based works have considered only static scenes or non-photorealistic human obstacles built from simulator assets, despite the importance of safe navigation in dynamic environments. To address these issues, we propose ReaDy-Go, a novel real-to-sim simulation pipeline that synthesizes photorealistic dynamic scenarios in target environments by augmenting a reconstructed static GS scene with dynamic human GS obstacles, and trains navigation policies using the generated datasets. The pipeline provides three key contributions: (1) a dynamic GS simulator that integrates static scene GS with a human animation module, enabling the insertion of animatable human GS avatars and the synthesis of plausible human motions from 2D trajectories, (2) a navigation dataset generation framework that leverages the simulator along with a robot expert planner designed for dynamic GS representations and a human planner, and (3) robust navigation policies to both the sim-to-real gap and moving obstacles. The proposed simulator generates thousands of photorealistic navigation scenarios with animatable human GS avatars from arbitrary viewpoints. ReaDy-Go outperforms baselines across target environments in both simulation and real-world experiments, demonstrating improved navigation performance even after sim-to-real transfer and in the presence of moving obstacles. Moreover, zero-shot sim-to-real deployment in an unseen environment indicates its generalization potential. Project page: https://syeon-yoo.github.io/ready-go-site/.

In cooperative environments, such as in factories or assistive scenarios, it is important for a robot to communicate its intentions to observers, who could be either other humans or robots. A legible trajectory allows an observer to quickly and accurately predict an agent's intention. In adversarial environments, such as in military operations or games, it is important for a robot to not communicate its intentions to observers. An illegible trajectory leads an observer to incorrectly predict the agent's intention or delays when an observer is able to make a correct prediction about the agent's intention. However, in some environments there are multiple observers, each of whom may be able to see only part of the environment, and each of whom may have different motives. In this work, we introduce the Mixed-Motive Limited-Observability Legible Motion Planning (MMLO-LMP) problem, which requires a motion planner to generate a trajectory that is legible to observers with positive motives and illegible to observers with negative motives while also considering the visibility limitations of each observer. We highlight multiple strategies an agent can take while still achieving the problem objective. We also present DUBIOUS, a trajectory optimizer that solves MMLO-LMP. Our results show that DUBIOUS can generate trajectories that balance legibility with the motives and limited visibility regions of the observers. Future work includes many variations of MMLO-LMP, including moving observers and observer teaming.

Quadrotor unmanned aerial vehicles (UAVs) are increasingly deployed in complex missions that demand reliable autonomous navigation and robust obstacle avoidance. However, traditional modular pipelines often incur cumulative latency, whereas purely reinforcement learning (RL) approaches typically provide limited formal safety guarantees. To bridge this gap, we propose an end-to-end RL framework augmented with model-based safety mechanisms. We incorporate physical priors in both training and deployment. During training, we design a physics-informed reward structure that provides global navigational guidance. During deployment, we integrate a real-time safety filter that projects the policy outputs onto a provably safe set to enforce strict collision-avoidance constraints. This hybrid architecture reconciles high-speed flight with robust safety assurances. Benchmark evaluations demonstrate that our method outperforms both traditional planners and recent end-to-end obstacle avoidance approaches based on differentiable physics. Extensive experiments demonstrate strong generalization, enabling reliable high-speed navigation in dense clutter and challenging outdoor forest environments at velocities up to 7.5 m/s}.

Sampling-based motion planners (SBMPs) are widely used for robot motion planning with complex kinodynamic constraints in high-dimensional spaces, yet their serial computation design results in planning speeds that scale poorly with problem complexity. Recent efforts to parallelize SBMPs have achieved significant speedups in finding feasible solutions; however, they provide no guarantees of optimizing an objective function. We introduce Kino-PAX$^{+}$, a massively parallel kinodynamic SBMP with asymptotic near-optimal guarantees. Kino-PAX$^{+}$ builds a sparse tree of dynamically feasible trajectories by decomposing traditionally serial operations into three massively parallel subroutines. The algorithm focuses computation on the most promising nodes within local neighborhoods for propagation and refinement, enabling rapid improvement of solution cost. We prove that, while maintaining probabilistic $δ$-robust completeness, this focus on promising nodes ensures asymptotic $δ$-robust near-optimality. Our results show that Kino-PAX$^{+}$ finds solutions up to three orders of magnitude faster than existing serial methods and achieves lower solution costs

We address the problem of reactive motion planning for quadrotors operating in unknown environments with dynamic obstacles. Our approach leverages a 4-dimensional spatio-temporal planner, integrated with vision-based Safe Flight Corridor (SFC) generation and trajectory optimization. Unlike prior methods that rely on map fusion, our framework is mapless, enabling collision avoidance directly from perception while reducing computational overhead. Dynamic obstacles are detected and tracked using a vision-based object segmentation and tracking pipeline, allowing robust classification of static versus dynamic elements in the scene. To further enhance robustness, we introduce a backup planning module that reactively avoids dynamic obstacles when no direct path to the goal is available, mitigating the risk of collisions during deadlock situations. We validate our method extensively in both simulation and real-world hardware experiments, and benchmark it against state-of-the-art approaches, showing significant advantages for reactive UAV navigation in dynamic, unknown environments.

Mobile robots have become indispensable for exploring hostile environments, such as in space or disaster relief scenarios, but often remain limited to teleoperation by a human operator. This restricts the deployment scale and requires near-continuous low-latency communication between the operator and the robot. We present MOSAIC: a scalable autonomy framework for multi-robot scientific exploration using a unified mission abstraction based on Points of Interest (POIs) and multiple layers of autonomy, enabling supervision by a single operator. The framework dynamically allocates exploration and measurement tasks based on each robot's capabilities, leveraging team-level redundancy and specialization to enable continuous operation. We validated the framework in a space-analog field experiment emulating a lunar prospecting scenario, involving a heterogeneous team of five robots and a single operator. Despite the complete failure of one robot during the mission, the team completed 82.3% of assigned tasks at an Autonomy Ratio of 86%, while the operator workload remained at only 78.2%. These results demonstrate that the proposed framework enables robust, scalable multi-robot scientific exploration with limited operator intervention. We further derive practical lessons learned in robot interoperability, networking architecture, team composition, and operator workload management to inform future multi-robot exploration missions.

A central problem in self-supervised learning is preventing representation collapse. Most methods avoid it through global mechanisms, such as contrastive expansion, variance constraints, decorrelating dimensions, or enforcing certain output distributions. In this work, we study a different design: short-range repulsion. We introduce Hypersolid, a self-supervised objective that combines view alignment with local collision avoidance. Our method induces a latent geometry of compact, semantically aligned neighborhoods with low anisotropy. This geometry is especially effective for unsupervised clustering and fine-grained separation, although it comes at the cost of weaker transferability.

Insect neuroethology provides a compelling biological template for efficient autonomous navigation. We draw an analogy between the formal embodied AI visual point-goal navigation task and the ability of insects to discover, learn, and refine visually guided paths around obstacles between a discovered food location and their nest. We develop a novel integrative model of mushroom body and central complex, two insect brain structures, that have been implicated, respectively, in associative learning and path integration. We demonstrate the mushroom body learning triggered by collisions results in adaptive obstacle avoidance and consequently optimised paths to the goal, corroborating the hypothesis of recent behavioural work that an insect can learn continuously as they travel. The embodied insect-inspired model achieves success rates comparable to recent state-of-the-art models at many orders of magnitude less computational cost in the standardised Habitat point-goal navigation benchmark. Testing in a more realistic simulated environment validates its robustness to perturbations.

Recent years have witnessed increasing interest in extending large language models into agentic systems. While the effectiveness of agents has continued to improve, efficiency, which is crucial for real-world deployment, has often been overlooked. This paper therefore investigates efficiency from three core components of agents: memory, tool learning, and planning, considering costs such as latency, tokens, steps, etc. Aimed at conducting comprehensive research addressing the efficiency of the agentic system itself, we review a broad range of recent approaches that differ in implementation yet frequently converge on shared high-level principles including but not limited to bounding context via compression and management, designing reinforcement learning rewards to minimize tool invocation, and employing controlled search mechanisms to enhance efficiency, which we discuss in detail. Accordingly, we characterize efficiency in two complementary ways: comparing effectiveness under a fixed cost budget, and comparing cost at a comparable level of effectiveness. This trade-off can also be viewed through the Pareto frontier between effectiveness and cost. From this perspective, we also examine efficiency oriented benchmarks by summarizing evaluation protocols for these components and consolidating commonly reported efficiency metrics from both benchmark and methodological studies. Moreover, we discuss the key challenges and future directions, with the goal of providing promising insights.

Crop monitoring is essential for precision agriculture, but current systems lack high-level reasoning. We introduce a novel, modular framework that uses a Vision Language Model (VLM) to guide robotic task planning by actively querying heterogeneous data sources, including enriched RGB camera feeds and 2D semantic occupancy maps, interleaved with robotic action primitives. We contribute a comprehensive benchmark for short- and long-horizon crop monitoring tasks in monoculture and polyculture environments. Our results show that while zero-shot VLMs perform robustly for short-horizon tasks (achieving 87% success, comparable to human experts), success drops significantly to under 10% for complex long-horizon, multi-target tasks. Despite this decline, task completion rates remain above 76% under noiseless conditions. Critically, the system degrades when relying on noisy semantic maps, demonstrating a key limitation in current VLM context grounding for sustained robotic operations. This work offers a deployable framework and critical insights into VLM capabilities and shortcomings for complex agricultural robotics.

Multi-constraint planning involves identifying, evaluating, and refining candidate plans while satisfying multiple, potentially conflicting constraints. Existing large language model (LLM) approaches face fundamental limitations in this domain. Pure reasoning paradigms, which rely on long natural language chains, are prone to inconsistency, error accumulation, and prohibitive cost as constraints compound. Conversely, LLMs combined with coding- or solver-based strategies lack flexibility: they often generate problem-specific code from scratch or depend on fixed solvers, failing to capture generalizable logic across diverse problems. To address these challenges, we introduce the Scalable COde Planning Engine (SCOPE), a framework that disentangles query-specific reasoning from generic code execution. By separating reasoning from execution, SCOPE produces solver functions that are consistent, deterministic, and reusable across queries while requiring only minimal changes to input parameters. SCOPE achieves state-of-the-art performance while lowering cost and latency. For example, with GPT-4o, it reaches 93.1% success on TravelPlanner, a 61.6% gain over the best baseline (CoT) while cutting inference cost by 1.4x and time by ~4.67x. Code is available at https://github.com/DerrickGXD/SCOPE.

Multi-agent systems based on large language models, particularly centralized architectures, have recently shown strong potential for complex and knowledge-intensive tasks. However, central agents often suffer from unstable long-horizon collaboration due to the lack of memory management, leading to context bloat, error accumulation, and poor cross-task generalization. To address both task-level memory inefficiency and the inability to reuse coordination experience, we propose StackPlanner, a hierarchical multi-agent framework with explicit memory control. StackPlanner addresses these challenges by decoupling high-level coordination from subtask execution with active task-level memory control, and by learning to retrieve and exploit reusable coordination experience via structured experience memory and reinforcement learning. Experiments on multiple deep-search and agent system benchmarks demonstrate the effectiveness of our approach in enabling reliable long-horizon multi-agent collaboration.

World models have become central to autonomous driving, where accurate scene understanding and future prediction are crucial for safe control. Recent work has explored using vision-language models (VLMs) for planning, yet existing approaches typically treat perception, prediction, and planning as separate modules. We propose UniDrive-WM, a unified VLM-based world model that jointly performs driving-scene understanding, trajectory planning, and trajectory-conditioned future image generation within a single architecture. UniDrive-WM's trajectory planner predicts a future trajectory, which conditions a VLM-based image generator to produce plausible future frames. These predictions provide additional supervisory signals that enhance scene understanding and iteratively refine trajectory generation. We further compare discrete and continuous output representations for future image prediction, analyzing their influence on downstream driving performance. Experiments on the challenging Bench2Drive benchmark show that UniDrive-WM produces high-fidelity future images and improves planning performance by 7.3% in L2 trajectory error and 10.4% in collision rate over the previous best method. These results demonstrate the advantages of tightly integrating VLM-driven reasoning, planning, and generative world modeling for autonomous driving. The project page is available at https://unidrive-wm.github.io/UniDrive-WM.

Vision-language-action (VLA) models provide strong action priors for robotic manipulation, but their reactive behavior can fail under distribution shift and long-horizon task structure. Recent VLA-guided planning methods improve execution by using pretrained policies to guide tree search, yet node selection still depends heavily on policy priors and visit-count exploration. Consequently, when the policy favors poor actions, the planner lacks a learned value signal to correct this bias. Prior work has shown that VLA representations encode rollout success and failure information, suggesting that they may also support value estimation during planning. We introduce Value-Guided Vision-Language-Action Planning and Search (V-VLAPS), which augments VLA-guided planning with a lightweight value head trained on offline VLA rollouts to predict Monte Carlo returns. These predictions guide Monte Carlo Tree Search in simulation toward higher-value branches. Across five LIBERO suites, V-VLAPS matches value-free planning baseline at the default search budget in aggregate, and analysis shows that many hard failures are root-level timeouts where predicted values are weakly separated. With a larger search budget, V-VLAPS improves over the baseline in all task suites with +6 percentage points on LIBERO-Object and +4 percentage points on LIBERO-10. Our results suggest that VLA representations can support not only failure prediction, but also value-guided planning when search reaches branches where value-based ranking matters.

Relative localization is essential for cooperation in autonomous multi-robot systems. Existing approaches either rely on shared environmental features or inertial assumptions, or they degrade under pairwise non-line-of-sight conditions and outliers in complex environments. Robustly and efficiently fusing inter-robot bearings, distances, and inertial measurements for tens of robots remains challenging. We present CREPES-X (Cooperative RElative Pose Estimation System with multiple eXtended features), a hierarchical relative localization framework that enhances speed, accuracy, and robustness under challenging conditions, without requiring any global information. The hardware packs infrared (IR) LEDs, an IR camera, an ultra-wideband module, and an IMU into a cube no larger than $6\,\text{cm}$ on each side. On this hardware, a two-stage hierarchical estimator meets different latency, accuracy, and robustness requirements. The single-frame estimator returns instantaneous relative poses from a closed-form solution with bearing outlier rejection. The multi-frame estimator then refines these poses with IMU pre-integration under robocentric relative kinematics, using loosely- and tightly-coupled optimization. Extensive simulations and real-world experiments validate the effectiveness of CREPES-X, demonstrating robustness of up to $90\%$ bearing outliers, resilience in challenging conditions, and RMSE of $7.0\,\text{cm}$ and $2.2^\circ$ in real-world datasets.

Autonomous driving requires reliable collision avoidance in dynamic environments. Nonlinear Model Predictive Controllers (NMPCs) are suitable for this task, but struggle in time-critical scenarios requiring high frequency. To meet this demand, optimization problems are often simplified via linearization, narrowing the horizon window, or reduced temporal nodes, each compromising accuracy or reliability. This work presents the first general convex obstacle avoidance formulation, enabled by a novel approach to integrating logic. This facilitates the incorporation of an obstacle avoidance formulation into convex MPC schemes, enabling a convex optimization framework with substantially improved computational efficiency relative to conventional nonconvex methods. A key property of the formulation is that obstacle avoidance remains effective even when obstacles lie outside the prediction horizon, allowing shorter horizons for real-time deployment. In scenarios where nonconvex formulations are unavoidable, the proposed method meets or exceeds the performance of representative nonconvex alternatives. The method is evaluated in autonomous vehicle applications, where system dynamics are highly nonlinear.

Many aerial tasks involving quadrotors demand both instant reactivity and long-horizon planning for obstacle avoidance, energy efficiency, or trajectory tracking. High-fidelity models enable accurate control but are too slow for long horizons. Low-fidelity planners scale but cannot directly control the system, necessitating cascaded architectures. Prevailing hierarchical approaches plan with a simplified model and use a high-fidelity controller for tracking, yet this decomposition is inherently suboptimal. The controller is limited by the coarse plan, and conventional MPC alternatives shorten the horizon to stay real-time feasible. We present UNIQUE, an MPC architecture that replaces this hierarchical stacking with temporal cascading. The planning problem is formulated as the second-tail horizon of a single multi-phase MPC, rather than being solved separately. We align costs across horizons, derive feasibility constraints for the point-mass planning model, and introduce transition constraints that convert high-fidelity states into meaningful low-fidelity states. Parallel point-mass and mixed-integer solvers address nonconvexities while incorporating progressive 3D obstacle smoothing over the planning horizon. In simulations and real flights, under equal computational budgets, UNIQUE improves closed-loop tracking by up to 75% compared with standard MPC and hierarchical baselines. Ablations and Pareto analyses confirm performance gains across variations in horizon, constraint approximations, and smoothing schedules.

Vision Language Models (VLMs) have shown promising planning capabilities, yet their success remains confined to the text domain, leaving visual decision-making relatively underexplored. Addressing this gap, we introduce Corrective Sequence Planning (CoSPlan) benchmark, where VLMs must plan a sequence of visual actions from an initial scene to a target scene. CoSPlan evaluates models on their ability to imagine and execute a coherent set of visual steps required to reach the goal (Step Completion). To prevent any shortcuts that simply describe the final scene, we introduce an erroneous action in decision-making, which must be detected (Error Detection) and corrected to reach the goal, enabling a deeper understanding of the task. CoSPlan spans across 4 tasks: maze navigation, block re-arrangement, image reconstruction, and object re-organization. Despite using advanced reasoning strategies such as Chain-of-Thought and Scene Graphs, VLMs struggle on CoSPlan, while still showing promising performance in the text domain. Addressing this, we propose Scene Graph Incremental updates (SGI), a novel training-free method to transform images into `textual' scene graphs, enabling step-by-step reasoning through iterative scene graph refinement. SGI yields an average of ~4.4% improvement on CoSPlan w/ generalization on PlanBench and VQA. Link for solving puzzles on the project page.

Monte Carlo Tree Search is a cornerstone algorithm for online planning, and its root-parallel variant is widely used when wall clock time is limited but best performance is desired. In environments with continuous action spaces, how to best aggregate statistics from different threads is an important yet underexplored question. In this work, we introduce a method that uses Gaussian Process Regression to obtain value estimates for promising actions that were not trialed in the environment. We perform a systematic evaluation across 6 different domains, demonstrating that our approach outperforms existing aggregation strategies while requiring a modest increase in inference time.

Reactive control can gracefully coordinate the motion of the base and the arm of a mobile manipulator. However, incorporating an accurate representation of the environment to avoid obstacles without involving costly planning remains a challenge. In this work, we present ReMoSPLAT, a reactive controller based on a quadratic program formulation for mobile manipulation that leverages a Gaussian Splat representation for collision avoidance. By integrating additional constraints and costs into the optimisation formulation, a mobile manipulator platform can reach its intended end effector pose while avoiding obstacles, even in cluttered scenes. We investigate the trade-offs of two methods for efficiently calculating robot-obstacle distances, comparing a purely geometric approach with a rasterisation-based approach. Our simulation experiments on both synthetic and real-world scans demonstrate the feasibility of the proposed method, achieving performance comparable to controllers that rely on perfect ground-truth information. We further validate the approach on a real robot platform more details: https://remosplat.github.io

Vision-Language-Action (VLA) models have demonstrated remarkable capabilities in generalizing across diverse robotic manipulation tasks. However, deploying these models in unstructured environments remains challenging due to the critical need for simultaneous task compliance and safety assurance, particularly in preventing potential collisions during physical interactions. In this work, we introduce a Vision-Language-Safe Action (VLSA) architecture, named AEGIS, which contains a plug-and-play safety constraint (SC) layer formulated via control barrier functions. AEGIS integrates directly with existing VLA models to improve safety with theoretical guarantees, while maintaining their original instruction-following performance. To evaluate the efficacy of our architecture, we construct a comprehensive safety-critical benchmark SafeLIBERO, spanning distinct manipulation scenarios characterized by varying degrees of spatial complexity and obstacle intervention. Extensive experiments demonstrate the superiority of our method over state-of-the-art baselines. Notably, AEGIS achieves over 50% improvement in obstacle avoidance rate while substantially increasing the task success rate by nearly 10%. All benchmark datasets, code, and supplementary materials are publicly available at https://vlsa-aegis.github.io/.

Autonomous driving planning systems perform nearly perfectly in routine scenarios using lightweight, rule-based methods but still struggle in dense urban traffic, where lane changes and merges require anticipating and influencing other agents. Modern motion predictors offer highly accurate forecasts, yet their integration into planning is mostly rudimental: discarding unsafe plans. Similarly, end-to-end models offer a one-way integration that avoids the challenges of joint prediction and planning modeling under uncertainty. In contrast, game-theoretic formulations offer a principled alternative but have seen limited adoption in autonomous driving. We present Bayesian Iterative Best Response (BIBeR), a framework that unifies motion prediction and game-theoretic planning into a single interaction-aware process. BIBeR is the first to integrate a state-of-the-art predictor into an Iterative Best Response (IBR) loop, repeatedly refining the strategies of the ego vehicle and surrounding agents. This repeated best-response process approximates a Nash equilibrium, enabling bidirectional adaptation where the ego both reacts to and shapes the behavior of others. In addition, our proposed Bayesian confidence estimation quantifies prediction reliability and modulates update strength, more conservative under low confidence and more decisive under high confidence. BIBeR is compatible with modern predictors and planners, combining the transparency of structured planning with the flexibility of learned models. Experiments show that BIBeR achieves an 11% improvement over state-of-the-art planners on highly interactive interPlan lane-change scenarios, while also outperforming existing approaches on standard nuPlan benchmarks.

Video generation has achieved remarkable progress in visual fidelity and controllability, enabling conditioning on text, layout, or motion. Among these, motion control - specifying object dynamics and camera trajectories - is essential for composing complex, cinematic scenes, yet existing interfaces remain limited. We introduce LAMP that leverages large language models (LLMs) as motion planners to translate natural language descriptions into explicit 3D trajectories for dynamic objects and (relatively defined) cameras. LAMP defines a motion domain-specific language (DSL), inspired by cinematography conventions. By harnessing program synthesis capabilities of LLMs, LAMP generates structured motion programs from natural language, which are deterministically mapped to 3D trajectories. We construct a large-scale procedural dataset pairing natural text descriptions with corresponding motion programs and 3D trajectories. Experiments demonstrate LAMP's improved performance in motion controllability and alignment with user intent compared to state-of-the-art alternatives establishing the first framework for generating both object and camera motions directly from natural language specifications. Code, models and data are available on our project page.

Multi-Agent Path Finding (MAPF) studies how to coordinate multiple agents to reach their goals without collisions and underpins a range of large-scale robotic systems, including automated warehousing and manufacturing. Recent advances enable MAPF solvers to compute high-quality plans for hundreds of agents. However, these plans are generated using simplified robot models with discretized time and action spaces. When they are deployed in physical systems, heterogeneous robot dynamics, asynchronous interactions, communication delays, and other real-world factors can lead to substantial deviations from planned performance. We bridge the gap between discrete planning and real-world execution through ExecTimeNet, a learned world model of MAPF execution that predicts how a discrete MAPF solution will unfold on physical robots, mapping each discrete action to its realized execution state, including its wall-clock completion time and the kinodynamic state in which it ends. Building on this capability, we first propose REMAP, an execution-aware MAPF framework that integrates execution-time estimation into planning, guiding the search toward MAPF solutions with improved execution performance. We also introduce ESADG, a post-planning optimization procedure that optimizes the execution schedule of a given MAPF solution while preserving path feasibility. We evaluate proposed frameworks in high-fidelity simulation with up to 300 agents and on physical robots. In simulation, ExecTimeNet predicts the execution state accurately and transfers to unseen maps and agent counts. Across simulation benchmarks spanning diverse map topologies, REMAP reduces delays by up to 21% over baselines, while ESADG achieves up to 40% normalized improvement. On physical hardware, the full pipeline reduces total execution time by up to 15.3%, demonstrating effective transfer from simulation to real-world deployment.

Egocentric video understanding requires procedural reasoning under partial observability and continuously shifting viewpoints. Current multimodal large language models (MLLMs) struggle with this setting, often generating plausible but visually inconsistent or weakly grounded responses. We introduce $\textbf{EgoVITA}$, a framework that decomposes egocentric video reasoning into a structured $\textit{plan-then-verify}$ process. The model first generates an $\textbf{egocentric plan}$: a causal sequence of anticipated actions from a first-person perspective. This plan is then evaluated by an $\textbf{exocentric verification}$ stage that uses third-person reasoning over the same video to verify its spatiotemporal and logical consistency, without exocentric video input. This decomposition enables cross-perspective feedback without requiring paired ego-exo supervision. To train this reasoning process, we adopt Group Relative Policy Optimization (GRPO) with two dense reward signals: one that grounds anticipated actions in subsequent visual observations and another that reinforces consistent third-person verification. $\textbf{EgoVITA}$ achieves state-of-the-art performance on egocentric reasoning benchmarks, outperforming Qwen2.5-VL-7B by $\mathbf{+7.7}$ on EgoBlind and $\mathbf{+4.4}$ on EgoOrient, while maintaining strong generalization on exocentric video tasks with only $52k$ training samples.

Model-based development accelerates prototyping, enables earlier experimentation, and ensures rigorous validation of system design intents. In multi-agent systems with complex asynchronous interactions and concurrency, formal verification, particularly model-checking, offers an automated means of confirming that desired properties hold. Timed Rebeca, an actor-based modelling language supporting reactive, concurrent, and timed behaviors, together with its model-checking tool, provides a powerful framework for this purpose. By leveraging these capabilities, Timed Rebeca can intuitively capture ROS2 node graphs, recurring physical signals, motion primitives, and other time-convertible behaviors. Nevertheless, modelling and verifying multi-robot systems entail significant challenges: abstracting intricate information, bridging the gap between discrete models and continuous system dynamics, and managing large state spaces while preserving fidelity. To address these challenges, we propose discretization strategies tailored to various data types and identify thresholds of abstraction that balance accuracy and tractability. We further introduce optimization techniques to accelerate verification. Our work demonstrates how to systematically design and verify multi-robot systems through Timed Rebeca, efficiently transform continuous dynamics into discrete models for model-checking, and maintain a practical, bidirectional flow between the abstract model and the ROS2 implementation. The accompanying Rebeca and ROS2 codebases, made openly available, serve as a foundational reference for researchers and developers aiming to model and verify advanced autonomous robotic systems.

Efficient routing of mobile robot fleets requires roadmaps with high redundancy, short path lengths, and sufficient node and edge clearance for conflict-free operation. Existing grid-based methods sacrifice geometric fidelity and impose Manhattan-distance path length constraints, whereas existing continuous-space methods neglect minimum distance constraints and transport demand. This paper proposes a continuous-space roadmap generation method that addresses this gap by placing nodes at convex corner points of the free space and at station interaction points, discretizing free space via local grid expansion, enforcing minimum inter-node and node-edge distance constraints derived from robot dimensions, and applying transport demand-driven K-shortest path pruning. The method is evaluated across three intralogistics environments using two multi-agent pickup and delivery (MAPD) solvers against three baselines: a reaction-diffusion sampling method (GSRM), an 8-connected grid, and random sampling. Under Priority Inheritance with Backtracking (PIBT), the proposed method outperforms GSRM by 1.2-23.4 % at maximum fleet size, the grid by at least 9.1 %, and random sampling by more than 10.4 % across all environments, with a space-time A* solver confirming these results. It further attains near-optimal normalized path lengths of 1.03-1.05 and the highest inter-station connectivity at comparable roadmap complexity.

Safe multi-agent motion planning (MAMP) under task-induced constraints is a critical challenge in robotics. Many real-world scenarios require robots to navigate dynamic environments while adhering to manifold constraints imposed by tasks. For example, service robots must carry cups upright while avoiding collisions with humans or other robots. Despite recent advances in decentralized MAMP for high-dimensional systems, incorporating manifold constraints remains difficult. To address this, we propose a manifold-constrained Hamilton-Jacobi reachability (HJR) learning framework for decentralized MAMP. Our method solves HJR problems under manifold constraints to capture task-aware safety conditions, which are then integrated into a decentralized trajectory optimization planner. This enables robots to generate motion plans that are both safe and task-feasible without requiring assumptions about other agents' policies. Our approach generalizes across diverse manifold-constrained tasks and scales effectively to high-dimensional multi-agent manipulation problems. Experiments show that our method outperforms existing constrained motion planners and operates at speeds suitable for real-world applications. Video demonstrations are available at https://youtu.be/RYcEHMnPTH8 .

Current evaluations of Large Language Model (LLM) agents primarily emphasize task completion, often overlooking resource efficiency and adaptability. This neglects a crucial capability: agents' ability to devise and adjust cost-optimal plans in response to changing environments. To bridge this gap, we introduce CostBench, a scalable, cost-centric benchmark designed to evaluate agents' economic reasoning and replanning abilities. Situated in the travel-planning domain, CostBench comprises tasks solvable via multiple sequences of atomic and composite tools with diverse, customizable costs. It also supports four types of dynamic blocking events, such as tool failures and cost changes, to simulate real-world unpredictability and necessitate agents to adapt in real time. Evaluating leading open-sourced and proprietary models on CostBench reveals a substantial gap in cost-aware planning: agents frequently fail to identify cost-optimal solutions in static settings, with even GPT-5 achieving less than 75% exact match rate on the hardest tasks, and performance further dropping by around 40% under dynamic conditions. By diagnosing these weaknesses, CostBench lays the groundwork for developing future agents that are both economically rational and robust.

Large language models (LLMs) have demonstrated significant potential to accelerate scientific discovery as valuable tools for analyzing data, generating hypotheses, and supporting innovative approaches in various scientific fields. In this work, we investigate how LLMs can handle the transition from conceptual research ideas to well-structured research plans. Effective research planning not only supports scientists in advancing their research but also represents a crucial capability for the development of autonomous research agents. Despite its importance, the field lacks a systematic understanding of LLMs' research planning capability. To rigorously measure this capability, we introduce the Idea2Plan task and Idea2Plan Bench, a set of benchmarks built from ICML 2025 and Nature Mental Health papers released after major LLM training cutoffs. Each benchmark instance includes a research idea and a grading rubric capturing the key components of valid plans. We further propose Idea2Plan JudgeEval, a complementary benchmark to assess the reliability of LLM-based judges against expert annotations. Experimental results show that GPT-5 achieves the strongest performance on the benchmark, though substantial headroom remains for improvement. Our study provides new insights into LLMs' capability for research planning and lays the groundwork for future progress.

Learning coordinated behaviors for multi-robot systems from only a few demonstrations is difficult because temporal task dependencies and spatial trajectory generation are tightly coupled, which increases the hypothesis space and often yields unstable generalization in data-scarce regimes. We present DDACE, a structured few-shot learning framework that introduces a structural inductive bias by explicitly decoupling temporal coordination from spatial trajectory synthesis. Demonstrations are first processed via spectral clustering to extract coordination structure and form interaction graphs. A Temporal Graph Network predicts action dependencies and sequences, while Gaussian Process models generate progress-parameterized geometric trajectories that adapt to new start/goal configurations. This factorized design reduces hypothesis coupling and improves data efficiency for few-shot multi-robot coordination. Extensive simulation studies and real-robot experiments show that DDACE produces stable coordinated executions from a small number of demonstrations and improves trajectory consistency compared to end-to-end imitation baselines under limited data. Additional materials are available at https://sites.google.com/view/ddace.

Reinforcement learning (RL), while powerful and expressive, can often prioritize performance at the expense of safety. Yet safety violations can lead to catastrophic outcomes in real-world deployments. Control Barrier Functions (CBFs) offer a principled method to enforce dynamic safety -- traditionally deployed online via safety filters. While the result is safe behavior, the fact that the RL policy does not have knowledge of the CBF can lead to conservative behaviors. This paper proposes CBF-RL, a framework for generating safe behaviors with RL by enforcing CBFs in training. CBF-RL has two key attributes: (1) minimally modifying a nominal RL policy to encode safety constraints via a CBF term, (2) and safety filtering of the policy rollouts in training. Theoretically, we prove that continuous-time safety filters can be deployed via closed-form expressions on discrete-time roll-outs. Practically, we demonstrate that CBF-RL internalizes the safety constraints in the learned policy -- both enforcing safer actions and biasing towards safer rewards -- enabling safe deployment without the need for an online safety filter. We validate our framework through ablation studies on navigation tasks and on the Unitree G1 humanoid robot, where CBF-RL enables safer exploration, faster convergence, and robust performance under uncertainty, enabling the humanoid robot to avoid obstacles and climb stairs safely in real-world settings without a runtime safety filter.

Centralized trajectory optimization in the joint space of multiple robots allows access to a larger feasible space that can result in smoother trajectories, especially while planning in tight spaces. Unfortunately, it is often computationally intractable beyond a very small swarm size. In this paper, we propose Flow-Opt, a learning-based approach towards improving the computational tractability of centralized multi-robot trajectory optimization. Specifically, we reduce the problem to first learning a generative model to sample different candidate trajectories and then using a learned Safety-Filter(SF) to ensure fast inference-time constraint satisfaction. We propose a flow-matching model with a diffusion transformer (DiT) augmented with permutation invariant robot position and map encoders as the generative model. We develop a custom solver for our SF and equip it with a neural network that predicts context-specific initialization. The initialization network is trained in a self-supervised manner, taking advantage of the differentiability of the SF solver. We advance the state-of-the-art in the following respects. First, we show that we can generate trajectories of tens of robots in cluttered environments in a few tens of milliseconds. This is several times faster than existing centralized optimization approaches. Moreover, our approach also generates smoother trajectories orders of magnitude faster than competing baselines based on diffusion models. Second, each component of our approach can be batched, allowing us to solve a few tens of problem instances in a fraction of a second. We believe this is a first such result; no existing approach provides such capabilities. Finally, our approach can generate a diverse set of trajectories between a given set of start and goal locations, which can capture different collision-avoidance behaviors.

In our deployed travel-planning service, most users give minimal inputs or free-form requests rather than the structured constraint checklists assumed by existing benchmarks. We therefore present TripScore, a behavior-grounded benchmark and evaluation framework built from real user logs and calibrated against 1,468 pairwise judgments by 203 travel experts. TripScore couples a hierarchical feasibility gate (format and commonsense) with a unified, point-wise reward that aggregates soft quality and preference fulfillment. Using TripScore as both evaluator and reward signal, we benchmark direct prompting, test-time compute, neuro-symbolic solvers, code agents, and fine-tuning. We find that reinforcement learning fine-tuning (e.g., GRPO) provides consistent gains over other approaches under the same base model and practical latency.

Data-driven planar pushing methods have recently gained attention as they reduce manual engineering effort and improve generalization compared to analytical approaches. However, most prior work targets narrow capabilities (e.g., side switching, precision, or single-task training), limiting broader applicability. We present a model-based framework for non-prehensile tabletop pushing that uses a single learned model to address multiple tasks without retraining. Our approach employs a recurrent GRU-based architecture with additional non-linear layers to capture object-environment dynamics while ensuring stability. A tailored state-action representation enables the model to generalize across uncertain dynamics, variable push lengths, and diverse tasks. For control, we integrate the learned dynamics with a sampling-based Model Predictive Path Integral (MPPI) controller, which generates adaptive, task-oriented actions. This framework supports side switching, variable-length pushes, and objectives such as precise positioning, trajectory following, and obstacle avoidance. Training is performed in simulation with domain randomization to support sim-to-real transfer. We first evaluate the architecture through ablation studies, showing improved prediction accuracy and stable rollouts. We then validate the full system in simulation and real-world experiments using a Franka Panda robot with markerless tracking. Results demonstrate high success rates in precise positioning under strict thresholds and strong performance in trajectory tracking and obstacle avoidance. Moreover, multiple tasks are solved simply by changing the controller's objective function, without retraining. While our current focus is on a single object type, we extend the framework by training on wider push lengths and designing a balanced controller that reduces the number of steps for longer-horizon goals.

This paper presents a framework to prompt multi-robot teams with high-level tasks using natural language expressions. Our objective is to use the reasoning capabilities of language models in understanding and decomposing multi-robot collaboration and decision-making tasks, but in settings where such models cannot be called at deployment time. However, it is hard to specify the behavior of an individual robot from a team instruction, and have it continuously adapt to actions from other robots. This necessitates a framework with the representational capacity required by the logic and semantics of a task, and yet supports decentralized, real-time operation. We solve this dilemma by recognizing that a task can be represented as a deterministic finite automaton, and that recurrent neural networks (RNNs) can encode numerous automata. This allows us to distill the logic and sequential decompositions of sub-tasks obtained from a language model into an RNN, and align its internal states with the semantics of a given task. This leads to a tiny model that encapsulates the reasoning of the language model and can be implemented onboard. To interpret the internal state of the RNN for a decentralized execution, we train a graph neural network control policy conditioned on the hidden states of the RNN and the language embeddings. We present evaluations on simulated and real-world multi-robot tasks that require sequential and collaborative behavior by the team, demonstrating scalable, robust, real-time performance -- sites.google.com/view/prompting-teams.

Co-optimizing safety and performance in large-scale multi-agent systems remains a fundamental challenge. Existing approaches based on multi-agent reinforcement learning (MARL), safety filtering, or Model Predictive Control (MPC) either lack strict safety guarantees, suffer from conservatism, or fail to scale effectively. We propose MAD-PINN, a decentralized physics-informed machine learning framework for solving the multi-agent state-constrained optimal control problem (MASC-OCP). Our method leverages an epigraph-based reformulation of SC-OCP to simultaneously capture performance and safety, and approximates its solution via a physics-informed neural network. Scalability is achieved by training the SC-OCP value function on reduced-agent systems and deploying them in a decentralized fashion, where each agent relies only on local observations of its neighbours for decision-making. To further enhance safety and efficiency, we introduce an Hamilton-Jacobi (HJ) reachability-based neighbour selection strategy to prioritize safety-critical interactions, and a receding-horizon policy execution scheme that adapts to dynamic interactions while reducing computational burden. Experiments on multi-agent navigation tasks demonstrate that MAD-PINN achieves superior safety-performance trade-offs, maintains scalability as the number of agents grows, and consistently outperforms state-of-the-art baselines.

Travel planning (TP) agent has recently worked as an emerging building block to interact with external tools/resources for travel itinerary generation, ensuring an enjoyable user experience. Despite its benefits, existing studies rely on hand-craft prompt and fixed agent workflow, hindering more flexible and autonomous TP agents. This paper proposes DeepTravel, an end-to-end agentic reinforcement learning framework for building an autonomous travel planning agent, capable of autonomously planning, executing tools, and reflecting on tool responses to explore, verify, and refine intermediate actions in multi-step reasoning. To achieve this, we first construct a robust travel sandbox by caching transportation, accommodation and POI data, facilitating TP agent training without being constrained by real-world APIs limitations (e.g., inconsistent outputs). Moreover, we develop a hierarchical reward modeling system, where a trajectory-level verifier first checks spatiotemporal feasibility and filters unsatisfied travel itinerary, and then the turn-level verifier further validate itinerary's detail consistency with tool responses, enabling efficient and precise reward service. Finally, we propose the reply-augmented reinforcement learning method that enables TP agent to periodically replay from a failure experience buffer, emerging notable agentic capacity. We deploy the trained TP agent in the DiDi Enterprise Solutions application. A three-month online test shows that it achieves 82% accuracy in travel itinerary generation. Comprehensive offline evaluations further demonstrate that DeepTravel enables small-sized LLMs (e.g., Qwen3-32B) to significantly outperform frontier LLMs (e.g., OpenAI o1/o3 and DeepSeek-R1) and existing TP agent frameworks.

Large Language Models (LLMs) have recently shown strong promise for robotic task planning, particularly through automatic planning domain generation. However, prior approaches largely treat generated planning domains as planning utilities, which are brittle under imperfect logical states and perception noise, overlooking their potential as scalable sources of reasoning supervision and structured reward signals. At the same time, reasoning LLMs depend on chain-of-thought (CoT) supervision that is expensive to collect for robotic tasks, and reinforcement learning (RL) faces challenges in reward engineering. We propose Self-CriTeach, an LLM self-teaching and self-critiquing framework in which an LLM autonomously generates symbolic planning domains that serve a dual role: (1) enabling large-scale generation of robotic planning problem-plan pairs, and (2) providing structured reward functions. First, the self-written domains enable large-scale generation of symbolic task plans, which are automatically transformed into extended CoT trajectories for supervised fine-tuning. Second, the self-written domains are reused as structured reward functions, providing dense feedback for reinforcement learning without manual reward engineering. This unified training pipeline yields a planning-enhanced LLM with higher planning success rates, stronger cross-task generalization, reduced inference cost, and resistance to imperfect logical states. GitHub Page: https://markli1hoshipu.github.io/Plan_LLM/

Multi-agent games in dynamic nonlinear settings are challenging due to the time-varying interactions among the agents and the non-stationarity of the (potential) Nash equilibria. In this paper we consider model-free games, where agent transitions and costs are observed without knowledge of the transition and cost functions that generate them. We propose a novel distributed policy structure that follows the communication constraints in multi-team games, with multiple agents per team, and learned through policy gradients. Our formulation is inspired by the structure of distributed policies in linear quadratic games, which take the form of time-varying linear feedback gains. In the nonlinear case, we model the policies as nonlinear feedback gains, parameterized by self-attention layers to account for the time-varying multi-agent communication topology. We demonstrate that our approach achieves strong performance in several settings, including distributed linear and nonlinear regulation, and simulated and real multi-robot pursuit-and-evasion games.

Efficient resource allocation in multi-agent systems requires autonomous agents to coordinate their decisions while balancing system-wide objectives with individual costs. This becomes increasingly challenging over long time horizons, where decisions that improve the current allocation may compromise future resource allocation, while decentralized agents have limited observations of the overall system. Multi-agent reinforcement learning (MARL) can learn such long-term dependencies via local observations, but directly applying it to large-scale coordination leads to rapidly growing decision spaces and inefficient training. To this end, we propose Hierarchical Reinforcement and Collective Learning (HRCL), a hierarchical framework that uses MARL to guide, rather than replace, decentralized multi-agent coordination. At the high level, MARL learns strategies that restrict the alternatives considered during coordination and guide agents in balancing system-wide and individual objectives. At the low level, agents perform efficient decentralized coordination under this strategic guidance. This separation reduces the learning space and allows short-term coordination trade-offs to be evaluated according to their long-term effects. Experiments on a synthetic benchmark show that HRCL converges substantially faster than standalone MARL and reduces system-wide and individual costs by 35.53% and 27.05%, respectively. Evaluations on energy self-management and drone swarm sensing further show improved resource allocation, power-peak regulation, and sensing efficiency. These results show that learning strategic guidance for an existing coordination process can retain scalable decentralized coordination without letting short-term decisions compromise future resource allocation.