Bachelorarbeit Informatik von Hiba Al-Najmi: The Impact Of Automation Levels On Trust And Acceptance Of Smart Home Warning Systems

Abstract: In this study, we address the under-explored dimension of human interaction with smart home warning systems during crisis scenarios, with focus on the identification of the optimal automation level. An experimental study was conducted with a sample of 51 participants in a controlled laboratory environment, which emulated smart home warning systems. Each participant engaged with one of three distinct levels of automation and encountered four distinct warnings with varying levels of danger. Subsequently, participants made decisions and potentially took actions in response to the warnings displayed by the system and the suggested protocols proposed by the system. These decisions encompassed choices to permit the system to proceed with execution, to halt or even revert the process, or to manually execute or refrain from executing the suggested actions. For instance, this could involve actions such as closing doors and windows in response to a gas leak incident originating from a nearby factory. Data were gathered via surveys and subsequently analyzed using ANOVA to discern significant differences in responses across the automation levels. The study revealed that the System Usability Scale (SUS) scores were consistently high across different levels of automation, indicating a uniformly high ease of use regardless of automation degree. Additionally, it was found that users responded significantly differently to critical versus non-critical smart home warnings, demonstrating the system’s effectiveness in appropriately categorizing and conveying the urgency of notifications.

Betreuer*in: Steffen Haesler, M.Sc.

Prüfer: Prof. Dr. Dr. Christian Reuter