Generalization to unseen environments remains a fundamental challenge in Vision-Language Navigation. To tackle this issue, we propose a novel framework that leverages world knowledge embedded within Multimodal Large Language Models. We introduce Collaborative Agents in Visual-Language Navigation (CA-VLN), a framework based on a dual-agent architecture. This architecture comprises a Knowledge Agent, which infuses the action prediction process with semantic context and commonsense reasoning, and a Hierarchical History Agent, which constructs a detailed episodic memory to enable long-horizon planning. The collaboration between these agents facilitates a dynamic interplay between high-level semantic understanding and grounded episodic experience. Extensive experiments on the R2R, REVERIE and SOON datasets demonstrate that our model achieves state-of-the-art performance, significantly improving generalization and navigation success in previously unobserved environments.

Vision-and-language navigation (VLN) traditionally relies on explicit reasoning chains, which, despite being interpretable, impose severe constraints on inference efficiency and scalability in long-range environments. Existing multimodal large language models (MLLMs) frequently encounter latency bottlenecks due to the generation of verbose textual narratives during decision-making. To address these limitations, we propose spatial reasoning vision-and-language navigation (SR-VLN), a novel framework that shifts the paradigm from explicit chain-of-thought (CoT) to an implicit spatial representation space. SR-VLN introduces a pyramidal hierarchical history framework integrated with perceptual compression to condense historical trajectories into multi-scale representations, effectively minimizing token overhead while preserving critical spatial semantics. Rather than generating verbose textual reasoning steps, SR-VLN employs compact, learnable spatial tokens (S-Tokens) to perform agile inference directly within the latent feature space. To establish robust causal mappings between these implicit states and navigational actions, we employ a hybrid training strategy that combines sparse reward supervision with reinforcement learning via GRPO. Extensive evaluations on the R2R, REVERIE, and SOON datasets demonstrate that SR-VLN achieves state-of-the-art overall navigation performance, while maintaining a comparable balance between accuracy and efficiency. Compared to explicit reasoning baselines, our method reduces token consumption by 68% and achieves a 4.1× speedup in inference while reaching a 76.02% success rate and a 73.80% SPL on the R2R unseen split, thereby facilitating near-real-time action prediction in long-range navigation environments.