RNR-Nav: A Real-World Visual Navigation System Using Renderable Neural Radiance Maps

TL;DR

This paper introduces RNR-Nav, a real-world visual navigation system that enhances neural radiance maps for better localization and navigation robustness in real environments, achieving significant performance improvements.

Contribution

The paper presents RNR-Map++, an improved neural radiance map with strategies to reduce information loss, and integrates it with a particle filter for robust real-time localization in real-world robots.

Findings

RNR-Nav achieves an 84.4% success rate in real-world navigation tasks.

RNR-Map++ significantly improves rendering quality and localization robustness.

The system outperforms previous methods by 68.8% in success rate.

Abstract

We propose a novel visual localization and navigation framework for real-world environments directly integrating observed visual information into the bird-eye-view map. While the renderable neural radiance map (RNR-Map) shows considerable promise in simulated settings, its deployment in real-world scenarios poses undiscovered challenges. RNR-Map utilizes projections of multiple vectors into a single latent code, resulting in information loss under suboptimal conditions. To address such issues, our enhanced RNR-Map for real-world robots, RNR-Map++, incorporates strategies to mitigate information loss, such as a weighted map and positional encoding. For robust real-time localization, we integrate a particle filter into the correlation-based localization framework using RNRMap++ without a rendering procedure. Consequently, we establish a real-world robot system for visual navigation utilizing RNR-Map++, which we call "RNR-Nav." Experimental results demonstrate that the proposed methods significantly enhance rendering quality and localization robustness compared to previous approaches. In real-world navigation tasks, RNR-Nav achieves a success rate of 84.4%, marking a 68.8% enhancement over the methods of the original RNR-Map paper.