Embodied Voice Assistant Markup Language

Abstract

Due to advances in robotics, the complexity of robots has increased greatly, allowing these devices to be used in increasingly challenging tasks. As a result, there is an increase in the complexity of developing programs for these robots, especially when using general purpose languages (GPLs). Even in simple programs this difficulty persists, as the algorithms need to communicate with sensors, obtaining and processing their values and, after processing, send some command to the robot’s actuators. In order to facilitate the programming of interactive sessions for social robotics platforms, this work proposes Embodied Voice Assistant Markup Language, a domain specific language based on XML, which can be applied to different social robots. The proposed language has elements for creating and manipulating variables, generating random numbers and conditional controls. In addition, it also proposes command abstractions for controlling multimodal interaction elements that are important in the human-robot interaction process. The Goal Question Metric paradigm (GQM) was used to structure the language assessment with 12 software developers, and then analyzed its clarity, effectiveness and perceived ease of use. This exploratory work presented very promising results, providing evidence that the proposed language is easy to use and understand.