X-Nav: Learning End-to-End Cross-Embodiment Navigation for Mobile Robots

2025-07-19

AI Summary

Existing navigation policies are typically tailored to specific robot morphologies, exhibiting poor generalization across diverse platforms. This paper introduces Nav-ACT, the first end-to-end cross-morphology navigation framework applicable to arbitrary wheeled and quadrupedal robots. Our method employs a two-stage learning paradigm: first, training morphology-aware expert policies in simulation using privileged information; second, distilling multi-morphology policy knowledge into a single unified policy via action chunking and a Transformer-based architecture. The resulting policy achieves zero-shot transfer—requiring no fine-tuning—to unseen morphologies and real-world environments. Experiments demonstrate that generalization performance improves with the number of training morphologies, and efficacy is validated across multiple physical robot platforms. The core contribution lies in breaking the morphology-coupling constraint, enabling truly morphology-agnostic, end-to-end, zero-shot navigation generalization.

Technology Category

• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Planning, Routing, and Scheduling: Learning for Planning and Scheduling

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• User Modeling, Personalization and Recommendation: Accountability, Transparency, and Ethics for personalization

Abstract

Existing navigation methods are primarily designed for specific robot embodiments, limiting their generalizability across diverse robot platforms. In this paper, we introduce X-Nav, a novel framework for end-to-end cross-embodiment navigation where a single unified policy can be deployed across various embodiments for both wheeled and quadrupedal robots. X-Nav consists of two learning stages: 1) multiple expert policies are trained using deep reinforcement learning with privileged observations on a wide range of randomly generated robot embodiments; and 2) a single general policy is distilled from the expert policies via navigation action chunking with transformer (Nav-ACT). The general policy directly maps visual and proprioceptive observations to low-level control commands, enabling generalization to novel robot embodiments. Simulated experiments demonstrated that X-Nav achieved zero-shot transfer to both unseen embodiments and photorealistic environments. A scalability study showed that the performance of X-Nav improves when trained with an increasing number of randomly generated embodiments. An ablation study confirmed the design choices of X-Nav. Furthermore, real-world experiments were conducted to validate the generalizability of X-Nav in real-world environments.

Problem

• Develop cross-embodiment navigation for diverse robots
• Train single policy for wheeled and quadrupedal robots
• Enable zero-shot transfer to unseen robot embodiments

Innovation

• End-to-end cross-embodiment navigation framework
• Transformer-based navigation action chunking (Nav-ACT)
• General policy from multi-expert distillation