Microblogging as a mechanism for human–robot interaction

Abstract

This paper presents a novel approach to social data analysis, exploring the use of microblogging to manage interaction between humans and robots, and presenting and evaluating an architecture that extends the use of social networks to connect humans and devices. The approach uses natural language processing – in the form of simple grammar-based techniques – to extract features of interest from textual data retrieved from a microblogging platform in real-time and generate appropriate executable code for the robot. The simple rule-based solution exploits some of the ‘natural’ constraints imposed by microblogging platforms to manage the potential complexity of the interactions and create bi-directional communication. In order to evaluate the developed system, an analysis of real-time, user-generated social media data is presented. The analysis demonstrates the feasibility of producing programmes from the social media data which lead to executable actions by a front-end application – an approach of immediate relevance to web-based systems, like question–answering engines, personal digital assistants, and smart home/office devices.