User-centric secured smart virtual assistants framework for disables

Abstract

Research on intelligent secured virtual assistant (ISVA) systems for disabled people is essential in order to meet the special requirements and overcome the difficulties they confront. The delicate nature of user interactions makes security and privacy considerations paramount in virtual assistant platforms. The gaps and weaknesses in existing systems can be identified by researching the context of current practice concerning their features, usability, limits in security procedures, and privacy restrictions. Therefore, we present a framework that combines blockchain-based security with federated learning (FL) to address the current shortcomings of virtual assistant technology. The examination focuses on two primary facets of cutting-edge virtual assistants. Firstly, it evaluates existing IoT-based virtual personal assistant systems designed for persons with disabilities, examining their features, usability, and limitations. The aim is to identify the specific needs and requirements of individuals with disabilities, considering their unique challenges and preferences in utilizing virtual assistant technologies. Second, considering the sensitivity of the information sent between users and virtual assistants, it explores the issues of security and privacy that arise while using such systems. The investigation covers authentication, data encryption, access control, and data privacy rules to provide a snapshot of the prevailing state protecting virtual assistants. Besides this, the framework strengthens the privacy and security of virtual assistants using blockchain technology. Through several empirical trials, it is found that the framework maintains better performance and usability, along with the provision of robust security mechanisms to safeguard user data and guarantee privacy.

Keywords

• Blockchain
• Authentication
• Privacy
• LSTM
• Global updation
• Operational efficiency