AURA: A Comprehensive Survey of AI-Based Assistants for the Visually Impaired

Abstract

Artificial intelligence has played a significant role in improving accessibility for visually impaired individuals by enabling real-time environmental understanding and multimodal interaction. This survey paper examines the design, methodology, and system components of AURA—an AI- powered virtual assistant developed to support individuals with visual impairments. The study evaluates various modules including speech processing, object detection, navigation, sound recognition, wearable technology, and guardian monitoring systems that collectively aim to provide a safe, intelligent, and accessible user experience. The survey further explores the challenges in current assistive solutions and how AURA bridges those gaps by integrating deep learning models, sensor-based feedback, multilingual communication, and real-time danger detection. The paper organizes the technical considerations into specific IEEE template-style sections while aligning them with the practical AI framework of AURA.

Introduction

AURA is an advanced assistive system designed to help visually impaired individuals overcome challenges in navigation, object recognition, and real-time awareness. Unlike traditional tools such as white canes or basic voice assistants, AURA integrates artificial intelligence, computer vision, speech processing, IoT sensors, and navigation algorithms to provide a comprehensive and real-time support system.

The system combines both hardware (smart guidance stick with sensors) and software (AI models and a guardian dashboard). It uses technologies like object detection, sound classification, face recognition, and multilingual communication to improve user independence and safety.

AURA is built using scalable frameworks such as Python, TensorFlow, PyTorch, and OpenCV, ensuring efficient and simultaneous processing of multiple inputs. It maintains system integrity through strict performance specifications, low-latency processing, and secure data handling.

The architecture relies on key components like GPS, APIs, and speech technologies (TTS/STT), along with precise units and equations for accurate sensing and navigation. It also addresses common issues in assistive systems—such as sensor errors and poor real-world performance—through advanced models and continuous user testing.

Finally, AURA emphasizes structured design, teamwork across multiple technical domains, and clear documentation using headings, figures, and tables to manage its complex, modular system effectively.

References

[1] D. Gurari, Q. Li, A. J. Stangl, A. Guo, C. Lin, K. Grauman, J. Luo, and J. P. Bigham, “VizWiz Grand Challenge: Answering visual questions from blind people,” in Proc. IEEE Conf. Computer Vision and Pattern Recognition (CVPR) Workshops, Salt Lake City, UT, USA, June 2018, pp. 1–10. [2] Microsoft Research, “Seeing AI: Making the visual world accessible,” Microsoft Research, Redmond, WA, USA, Tech. Rep., 2017. [3] Google Research, “Responsible development of Lookout for people with visual impairments,” Google LLC, Mountain View, CA, USA, White Paper, 2023. [4] B. J. Nguyen, J. P. Bigham, and A. J. Stangl, “Improvement in patient- reported quality-of-life outcomes in severely visually impaired individuals using the Aira assistive technology system,” Translational Vision Science & Technology, vol. 7, no. 5, pp. 1–9, Aug. 2018. [5] Be My Eyes Research Team, “Be My Eyes: Remote visual assistance for blind and low-vision users,” Be My Eyes, Copenhagen, Denmark, Industry Report, 2016. [6] D. Gurari and J. P. Bigham, “AI for accessibility: Opportunities and challenges,” ACM Computing Surveys, vol. 54, no. 6, pp. 1–36, Dec. 2022. [7] D. Gurari et al., “VizWiz datasets and challenges: Enabling visual question answering for blind people,” arXiv preprint arXiv:2002.08577, 2020.

Copyright

Copyright © 2026 Prof. Gauri Bobade, Chaitanya Jadhav, Alpha Londhe , Atharva Arote , Eshita Gangurde . This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.