Abstract
In order to improve the interaction between humans and robots, robots need to be able to move about in a way that is appropriate to the complex environments around them. One way to investigate how the robots should move is through the lens of theatre, which provides us with ways to analyze the robot's movements and the motivations for moving in particular ways. In particular, this has proven useful for improving robot navigation. By altering the costmaps used for path planning, robots can navigate around their environment in ways that incorporate additional contexts. Experimental results with user studies have shown altered costmaps to have a significant effect on the interaction, although the costmaps must be carefully tuned to get the desired effect. The new layered costmap algorithm builds on the established open-source navigation platform, creating a robust system that can be extended to handle a wide range of contextual situations.
Committee Chair
Bilge Mutlu
Committee Members
William D Smart, Robert Pless, Chris Gill, Caitlin Kelleher
Degree
Doctor of Philosophy (PhD)
Author's Department
Computer Science & Engineering
Document Type
Dissertation
Date of Award
Winter 12-15-2014
Language
English (en)
DOI
https://doi.org/10.7936/K77P8WHT
Recommended Citation
Lu, David Vincent, "Contextualized Robot Navigation" (2014). McKelvey School of Engineering Theses & Dissertations. 62.
The definitive version is available at https://doi.org/10.7936/K77P8WHT
Comments
Permanent URL: https://doi.org/10.7936/K77P8WHT