Search papers, labs, and topics across Lattice.
This paper introduces an imitation-learning-based framework for safe visual navigation that incorporates embodiment awareness, addressing the ambiguity in action prediction caused by differing embodiments. By constructing a cross-embodiment navigation dataset from Internet videos and utilizing embodiment geometry as conditional tokens, the approach reduces action ambiguity and enhances navigation performance. Experimental results indicate significant improvements in navigation across various embodiment settings, highlighting the effectiveness of the proposed method in real-world adaptation scenarios.
Action ambiguity in navigation is drastically reduced by leveraging embodiment geometry, leading to safer and more effective navigation policies across diverse agents.
Cross-embodiment navigation is a key challenge in embodied intelligence. Due to differences in embodiment, the same visual observation may imply different actions for different agents, making prediction ambiguous when relying solely on vision. Existing studies mainly rely on reinforcement learning, which requires large-scale interaction and careful reward design, making it difficult to support scalable pretraining and real-world adaptation. In contrast, imitation-learning-based approaches remain limited. To address these challenges, we propose an imitation-learning-based embodiment-aware navigation framework with a modular multi-stage design. In pretraining, we construct a cross-embodiment navigation dataset from Internet videos and introduce embodiment geometry as conditional tokens to reduce action ambiguity under the same observation. In fine-tuning, we design a multimodal information injection mechanism based on a decoupled architecture. Specifically, we design a trajectory augmentation strategy to generate high-risk samples, which are used to train spatial perception and risk-aware correction separately, thereby explicitly incorporating embodiment geometry for safe navigation. Experimental results show that the proposed method effectively improves navigation performance across different embodiment settings, demonstrating the effectiveness of incorporating embodiment geometry into embodied navigation.
