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

TL;DR

X-Nav introduces a unified end-to-end navigation framework that generalizes across various robot embodiments, trained via a two-stage learning process, and validated through extensive simulated and real-world experiments.

Contribution

The paper presents a novel cross-embodiment navigation framework using a two-stage learning process with transformer-based policy distillation, enabling zero-shot transfer across diverse robots.

Findings

X-Nav achieves zero-shot transfer to unseen robot embodiments.

Performance improves with more diverse training embodiments.

Real-world experiments validate generalization capabilities.

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.