RVN-Bench: A Benchmark for Reactive Visual Navigation

TL;DR

RVN-Bench is a new indoor visual navigation benchmark that emphasizes collision avoidance, supporting diverse environments and both online and offline learning, to improve safe robotic navigation.

Contribution

It introduces RVN-Bench, a comprehensive collision-aware benchmark for indoor visual navigation, with tools for training, evaluation, and dataset generation, filling a gap in existing benchmarks.

Findings

Policies trained on RVN-Bench generalize well to unseen environments.

RVN-Bench enables effective evaluation of collision-aware navigation strategies.

The benchmark supports both online reinforcement learning and offline dataset generation.

Abstract

Safe visual navigation is critical for indoor mobile robots operating in cluttered environments. Existing benchmarks, however, often neglect collisions or are designed for outdoor scenarios, making them unsuitable for indoor visual navigation. To address this limitation, we introduce the reactive visual navigation benchmark (RVN-Bench), a collision-aware benchmark for indoor mobile robots. In RVN-Bench, an agent must reach sequential goal positions in previously unseen environments using only visual observations and no prior map, while avoiding collisions. Built on the Habitat 2.0 simulator and leveraging high-fidelity HM3D scenes, RVN-Bench provides large-scale, diverse indoor environments, defines a collision-aware navigation task and evaluation metrics, and offers tools for standardized training and benchmarking. RVN-Bench supports both online and offline learning by offering an environment for online reinforcement learning, a trajectory image dataset generator, and tools for producing negative trajectory image datasets that capture collision events. Experiments show that policies trained on RVN-Bench generalize effectively to unseen environments, demonstrating its value as a standardized benchmark for safe and robust visual navigation. Code and additional materials are available at: https://rvn-bench.github.io/.