Robust Dialog Design for Emergency Situations with AI-Based Voice Assistants

David Feilacher; Sabine Lehner; Mario Heller; Berhard Girsule; Andreas Jakl

doi:10.60522/o:7211

Open in new window

Title (eng)

Robust Dialog Design for Emergency Situations with AI-Based Voice Assistants

Author

David Feilacher

St. Pölten University of Applied Sciences

Sabine Lehner

St. Pölten University of Applied Sciences

Mario Heller

St. Pölten University of Applied Sciences

Berhard Girsule

St. Pölten University of Applied Sciences

Andreas Jakl

Description (eng)

With rising life expectancy and seniors seeking to live independently for as long as possible, automated emergency detection and communication will play a vital role in ensuring quick emergency responses in the future. Voice assistants enable a quick and hands-free reaction by the users and can also take their feedback into account. This paper presents the methods employed by the Smart Companion project to improve the dialogue and technical structures of a voice assistant system for emergency situations. We propose a dialogue structure, guided by a user centered design process, that issues concise and closed-ended questions while being able to process a wide range of possible user responses. This structure can handle any verbal user input, including non-responses and unintelligible utterances, through feedback and fallback mechanisms. To avoid technical errors or unresponsive users prematurely terminating an emergency call, conversations can be restarted at any point in the dialogue structure, by working in tandem with the other Smart Companion components. This also enables the emergency system to proactively ascertain if the user needs help. The technical implementation provides an easy dialogue configuration, adaptation, and expansion, with a custom state machine design supervising the dialogue flow.

Keywords (eng)

Voice AssistantsHuman Machine CommunicationAIDialogue DesignState Machine Design

Subject (eng)

ÖFOS 2012 -- 5080 -- Media and Communication Sciences

Subject (eng)

ÖFOS 2012 -- 202022 -- Information technology

Subject (eng)

ÖFOS 2012 -- 202002 -- Audiovisual media

Subject (eng)

ÖFOS 2012 -- 6040 -- Arts

Type (eng)

conference object