Advancements in Assistive Technology: A Blind Aid Bot for Enhanced Spatial Awareness and Navigation
Tushar Mathur, Nandini Jain, Utkarsh Jaiswal, Rachita Kothiyal, Puneet Sharma, Rohit Bathla · 2024
Independence of people with vision impairments is necessary as they face many hurdles in their daily lives from locating objects to navigating through obstacles. In the current environment, it is still challenging to provide the visually impaired with appropriate guidance and to find solutions to their difficulties. In order to enhance the awareness and day-to-day activities of visually impaired users, this project introduces a software-centred approach called the visual assistant. The software architecture of the blind assistance robot is based on deep learning and can use patterns for object recognition, intervention and optimization. The system uses convolutional neural networks (CNN) and you only look once (yolov5) and is trained on many different urban and indoor environments to ensure robustness and accuracy in different situations. Integration with natural language processing (NLP) makes it intuitive, allowing users to receive messages and send commands with simple voice. Test in a simulated environment to check the effectiveness of algorithms and user interfaces. Preliminary results show significant improvements in reporting accuracy and user confidence when interacting in an unfamiliar environment. Responses from users in the first test show the potential of the blind program to reduce the daily problems of the visually impaired. This paper not only demonstrates the feasibility of software solutions in helping the visually impaired, but also opens up personal development opportunities. Future work will focus on real-world testing and integration of adaptive learning techniques to further improve user interaction and environment-based performance.