Raspberry Pi - based Alexa System

Abstract

Spurred by the advancement of artificial intelligence and natural language processing, virtual assistants such as Google Assistant and Alexa are becoming widely available. Despite the popularity of virtual assistants, significant research showed that voice-only assistants have poor usability. In particular, Alexa possesses limited flexibility to personalize experiences for users. Most existing Alexa skills in the market are only able to execute and reply to simple queries. Thus, this project aims to explore the implementation of Amazon Alexa on a Raspberry Pi and investigate how an Alexa Skill can be built to increase interactivity through the use of dialog management techniques and application of design principles. The Raspberry Pi based Alexa system is built on a Raspberry Pi Model B unit. Using the Alexa Voice Service (AVS), it enables communication to the various cloud services. The system was implemented mainly using Javascript to enable extensibility for further development of the project. In light of this, the proposed system comprises of three main modules: medicine reminder, facial recognition and voice journal. The proposed system aims to not only bring convenience to users but to provide a rich voice user experience.