Contextual Coaching: Support for Well-being based on Environmental Sensor Information and Conversational AI

Abstract

Recently, robots in home environments have become available and their use ranges from companions to service providers. In this paper well-being and healthy living are considered in smart home environments where a robot agent is available to provide friendly support for the users’ daily life and encourage them for active independent living. Environmental sensors provide real-time information about the temperature, humidity, CO2 level and light. We further use conversational AI to enable the robot to interact with the user in natural language concerning the environmental state, and accordingly provide recommendations and reminders related to the current environmental state. The robot agent can talk about environmental sensors and their impact on the user’s well-being. A crucial component in the framework is the use of knowledge graphs which store the information relevant for the conversation, and also allow inferences about the knowledge state. The interaction is demonstrated on the Furhat robot agent using Rasa Conversational AI.