Crowdsourcing Task Traces for Service Robotics

TL;DR

This paper addresses enabling end-user developers to script service-robot tasks by leveraging crowdsourced task traces. It introduces a lightweight web interface to collect decontextualized, step-by-step traces via MTurk, yielding 207 traces across 18 task categories. The authors discuss how these traces can form task models to guide end-user development tools, enabling suggestions for missing steps, loops, or branches. Findings indicate the interface scales and that crowdsourced traces are rich enough to inspire personalized task specifications, though more data and automated curation are needed.

Abstract

Demonstration is an effective end-user development paradigm for teaching robots how to perform new tasks. In this paper, we posit that demonstration is useful not only as a teaching tool, but also as a way to understand and assist end-user developers in thinking about a task at hand. As a first step toward gaining this understanding, we constructed a lightweight web interface to crowdsource step-by-step instructions of common household tasks, leveraging the imaginations and past experiences of potential end-user developers. As evidence of the utility of our interface, we deployed the interface on Amazon Mechanical Turk and collected 207 task traces that span 18 different task categories. We describe our vision for how these task traces can be operationalized as task models within end-user development tools and provide a roadmap for future work.

Figures (2)

• Figure 1: Our crowdsourcing interface collects individual task traces, which are organized by task category.
• Figure 2: The web interface that we created to collect task traces. The Layout of Home pane uses graphics from LimeZu.