[en] Virtual assistants such as Alexa, Siri, Cortana and Google Assistant increasingly find ways to enter our homes and everyday lives, where they serve as powerful human-machine interfaces to control devices through natural language. In this work, we explore the use of virtual assistants to control single components within an intelligent industrial robot assistant system for disassembly applications. Following a short introduction and an overview of commercially available virtual assistants, we present our system architecture, which integrates Amazon's Alexa through an Echo Dot device. Utilizing the Alexa Skill Kit, we build a voice user interface that contains dialogs to control various device functions and assistive system behaviors. By connecting Alexa Voice Service with Amazon’s Lambda and IoT web services, we can parametrize machine commands depending on the user's voice input and, through a Raspberry Pi, which routes messages between the internet and decoupled machine network, send them to the devices within the intelligent robot assistance system. With the reversed communication flow, we can update and store the current state of the devices in the Amazon Web Services IoT Shadow. Utilizing the IoT Shadow, we are able to inform the user upon request about the current state of the devices by synthesize speech with the Alexa Voice Services and vocalize it through the Echo Dot. One implemented assistive system behavior and the related dialog to activate the behavior is further explained in this work. In the conclusion, we address our gathered positive and negative experience to date.
[University of Luxembourg > Faculty of Science, Technology and Communication (FSTC) > >]
