Qingrong He

Tianle Zhang, Zhihao Yuan, Dafeng Chi et al.

Robotic autonomy in open-world environments is fundamentally limited by insufficient data diversity and poor cross-embodiment generalization. Existing robotic datasets are often limited in scale and task coverage, while relatively large differences across robot embodiments impede effective behavior knowledge transfer. To address these challenges, we propose JoyAI-RA, a vision-language-action (VLA) embodied foundation model tailored for generalizable robotic manipulation. JoyAI-RA presents a multi-source multi-level pretraining framework that integrates web data, large-scale egocentric human manipulation videos, simulation-generated trajectories, and real-robot data. Through training on heterogeneous multi-source data with explicit action-space unification, JoyAI-RA effectively bridges embodiment gaps, particularly between human manipulation and robotic control, thereby enhancing cross-embodiment behavior learning. JoyAI-RA outperforms state-of-the-art methods in both simulation and real-world benchmarks, especially on diverse tasks with generalization demands.

Kangyi Wu, Pengna Li, Kailin Lyu et al.

Vision-Language Navigation(VLN) requires an agent to navigate through 3D environments by following natural language instructions. While recent Video Large Language Models(Video-LLMs) have largely advanced VLN, they remain highly susceptible to State Drift in long scenarios. In these cases, the agent's internal state drifts away from the true task execution state, leading to aimless wandering and failure to execute essential maneuvers in the instruction. We attribute this failure to two distinct cognitive deficits: Progress Drift, where the agent fails to distinguish completed sub-goals from remaining ones, and Memory Drift, where the agent's history representations degrade, making it lose track of visited landmarks. In this paper, we propose a Dual-Anchoring Framework that explicitly anchors the instruction progress and history representations. First, to address progress drift, we introduce Instruction Progress Anchoring, which supervises the agent to generate structured text tokens that delineate completed versus remaining sub-goals. Second, to mitigate memory drift, we propose Memory Landmark Anchoring, which utilizes a Landmark-Centric World Model to retrospectively predict object-centric embeddings extracted by the Segment Anything Model, compelling the agent to explicitly verify past observations and preserve distinct representations of visited landmarks. Facilitating this framework, we curate two extensive datasets: 3.6 million samples with explicit progress descriptions, and 937k grounded landmark data for retrospective verification. Extensive experiments in both simulation and real-world environments demonstrate the superiority of our method, achieving a 15.2% improvement in Success Rate and a remarkable 24.7% gain on long-horizon trajectories. To facilitate further research, we will release our code, data generation pipelines, and the collected datasets.

Qingrong He, Kejun Lin, Shizhe Chen et al.

This work addresses the 3D situated reasoning task which aims to answer questions given egocentric observations in a 3D environment. The task remains challenging as it requires comprehensive 3D perception and complex reasoning skills. End-to-end models trained on supervised data for 3D situated reasoning suffer from data scarcity and generalization ability. Inspired by the recent success of leveraging large language models (LLMs) for visual reasoning, we propose LLM-TPC, a novel framework that leverages the planning, tool usage, and reflection capabilities of LLMs through a ThinkProgram-reCtify loop. The Think phase first decomposes the compositional question into a sequence of steps, and then the Program phase grounds each step to a piece of code and calls carefully designed 3D visual perception modules. Finally, the Rectify phase adjusts the plan and code if the program fails to execute. Experiments and analysis on the SQA3D benchmark demonstrate the effectiveness, interpretability and robustness of our method. Our code is publicly available at https://qingrongh.github.io/LLM-TPC/.

Xi Lin, Jiayi Li, Kangyi Wu et al.

Robots deployed in unstructured human environments must frequently execute long-horizon missions, such as find the mug, then the chair, then the printer, under strict operational constraints. While contemporary zero-shot Object Navigation (ObjectNav) agents leverage Vision-Language Models (VLMs) to effectively localize semantic targets, they operate as purely reactive systems that inherently lack global resource awareness. Consequently, these agents inadvertently exhaust critical budgets, including time and battery, on infeasible subgoals due to partial observability, failing to balance local exploration with global mission viability. To bridge this gap by injecting resource-rationality into the navigation loop, we present MORN (Metacognitive Object-goal Regulation Navigation), an executive architecture inspired by Dual-Process Theory in cognitive science. MORN augments frozen navigation backbones with a System 2 meta-controller that continuously monitors the System 1 locomotor. By formalizing three neuro-cognitive states, Potentiality Index, Persistence Gating, and Evidence Accumulation, MORN dynamically regulates the mission schedule based on online estimates of progress velocity and perceptual uncertainty. This mechanism effectively neutralizes the Sunk Cost Fallacy, enabling agents to abort zombie goals early and decisively commit to achievable ones. Extensive experiments on the HM3D dataset demonstrate that MORN improves Goal Completion Rate (CR) from 0.23 to 0.30 and reduces Wasted Step Fraction (WSF) from 0.90 to 0.70, establishing that in resource-constrained autonomy, the metacognitive awareness of global resources is as critical as the reactive ability to navigate.