Home › Categories › Technology› Navigation Aid

Navigation Aid.pdf

weiss20a.pdf

🔗 Source: proceedings.mlr.press

📊 Size: 4.09 MB

👤 Author: Anonymous Submission

⬇️ Downloads: 435

Summary

Millions of blind and visually-impaired people struggle to navigate urban environments, relying on smartphones for path-planning and white canes or guide dogs for local information. This work introduces SEVN, a sidewalk simulation environment and neural network-based approach to creating a navigation agent, containing panoramic images with labels for house numbers, doors, and street signs, and formulations for several navigation tasks. The policy model fuses multi-modal observations to navigate to a goal door, demonstrating strong performance on a street segment navigation task. The primary contributions include a benchmark dataset, an OpenAI Gym-compatible environment, and a novel neural architecture for RL trained with Proximal Policy Optimization (PPO) to fuse images, GPS, and scene text for navigation.

Download File

Description

Millions of blind and visually-impaired people struggle to navigate urban environments, relying on smartphones for path-planning and white canes or guide dogs...

Technical Information

File Format: PDF
File Size: 4.09 MB
Pages: 14
Language: EN
Author: Anonymous Submission
Total Downloads: 435
Last Updated: 2 weeks ago

Document Overview

This PDF document about Navigation Aid provides comprehensive information and guidance. Whether you're a beginner or advanced user, this resource offers valuable insights into Navigation Aid.