Modeling Error and Delayed Behavior During Interruptions in Human-Robot Interaction Tasks Using a Cognitive Model
- Bachelorarbeit -
Context and Motivation
In today's day and age, knowledge workers, especially students, are confronted with a myriad of distractions. One of the more recent phenomena is the emergence of AI assistants (e.g. robots) meant to help users with timely and useful suggestions. These suggestions can also cause disruption in the user's workflow and lead to delays and errors. To improve human-robot interaction, it is crucial to understand the mistakes that are likely to occur and how they affect performance. Generating relevant experimental results for this problem can be challenging because large sample sizes are required. Furthermore, conducting trials with human participants is costly and time-consuming.
Goal of the Thesis
This thesis aims to predict how the usefulness and timing of robot suggestions influence the type and frequency of user errors and also the delay in resuming the original task after an interruption. For this, a cognitive model is employed with the intention of producing human-like results in a specific task.
Approach
The experimental setup is based on a previously established method that receives minor alterations, extending the logging capabilities and manipulating how and when suggestions appear. ACT-R will be the cognitive model that simulates and predicts human error behavior and task resumption delays.
Research Question
How do timing and usefulness of interruptions influence the type and quantity of errors made and task resumption delay in a robot-assisted task?
Anforderungen/Kenntnisse:
Kognitive Modellierung mit ACT-R, Durchführung einer Evaluierung
Bearbeitung:
Tom Jonas Raddei
Betreuung:
Prof. Dr. rer.nat. Nele Rußwinkel
Institut für Informationssysteme
Ratzeburger Allee 160 ( Gebäude 64 - 2. OG)
23562 Lübeck
Telefon: 0451 / 3101 5700
