AI-based Chatbot for Physically Challenged People
Aashdeep Singh, P. Murugeswari, S D Prabu Ragavendiran, Amanpreet Kaur, Gurpreet Singh, Sathiyamoorthy Margabandu · 2022 International Conference on Edge Computing and Applications (ICECAA) · 2022
Blindness is a state where a person loses his/her ability to see completely or partially. It is also said to be one of the most widely found disabilities across the world. Most of the time, it is impossible to find a complete cure for this condition. Thus, people with blindness adapt themselves to living with this blindness as a part of them. To help visually impaired people, this study aims in developing a chatbot which uses artificial intelligence to predict the most relevant answers to the users’ questions. This bot is designed in such a way that it allows the user to send voice notes instead of texts like the conventional chatbots. The chatbot is developed based on a particular architecture. The user-sent text is first converted to text by this architecture. Deep neural networks and APIs that are used for speech-to-text and text-to-speech conversion make up the chatbot's main architecture. The optimal response is then determined for the user from this text, and it is sent to the user. The inner parts of the architecture consist of an AI algorithm named the deep neural networks, speech-to-text, and text-to-speech APIs, libraries like GTTS, etc. This chatbot can be trained to add more functions, such as the ability to automatically order essentials from the websites indicated, transmit emergency signals, and remind users to take their medications, eat their meals, and drink their water in the future.