Navigation Agents for the Visually Impaired: A Sidewalk Simulator and Experiments

TL;DR

This paper introduces SEVN, a sidewalk simulation environment with multimodal data, and demonstrates how reinforcement learning can be used to develop navigation agents to assist visually impaired individuals in urban environments.

Contribution

The work presents a new sidewalk simulation environment, SEVN, and explores reinforcement learning for navigation tasks tailored to visually impaired users.

Findings

RL agent using PPO can navigate using multimodal inputs

SEVN provides labeled panoramic images for navigation tasks

Experimental results establish a foundation for future BVI navigation research

Abstract

Millions of blind and visually-impaired (BVI) people navigate urban environments every day, using smartphones for high-level path-planning and white canes or guide dogs for local information. However, many BVI people still struggle to travel to new places. In our endeavor to create a navigation assistant for the BVI, we found that existing Reinforcement Learning (RL) environments were unsuitable for the task. This work introduces SEVN, a sidewalk simulation environment and a neural network-based approach to creating a navigation agent. SEVN contains panoramic images with labels for house numbers, doors, and street name signs, and formulations for several navigation tasks. We study the performance of an RL algorithm (PPO) in this setting. Our policy model fuses multi-modal observations in the form of variable resolution images, visible text, and simulated GPS data to navigate to a goal door. We hope that this dataset, simulator, and experimental results will provide a foundation for further research into the creation of agents that can assist members of the BVI community with outdoor navigation.