This research examined how information regarding a robot teammate's reliability and the consequences for mistakes made by a robot in its task influence reliance on the robot by a human teammate. Of interest in this research was the notion of appropriate reliance: relying on a robot teammate's decisions when it is performing well and not relying on its decisions when it is performing poorly. An experiment was conducted in which participants interacted with an autonomous robot teammate while performing a cordon and search operation within a virtual reality simulation environment. Participants were responsible for monitoring the perimeter of a search area while their robot teammate searched the area for target objects. The robot's reliability shifted between 90% (good) or 10% (poor) based on the environment it was currently searching. Participants were assigned to one of four experimental groups that differed according to: (a) the information they were provided about the robot teammate's reliability and which factors influenced it (minimal information or complete information), and (b) the specific consequences for the robot missing target objects during its search (low risk or high risk). Findings indicated that participants provided with complete reliability information relied more appropriately on the robot's decisions (i.e., participants relied more when the robot performed well and relied less when it was performing poorly) than participants who did not receive this information. Appropriate reliance was not, however, affected by the consequences for mistakes made on the robot's task. These results provide support for the notion that informing individuals of the factors influencing a robot's reliability helps them to rely more appropriately on its decisions.
| Identifer | oai:union.ndltd.org:ucf.edu/oai:stars.library.ucf.edu:etd-7852 |
| Date | 01 January 2019 |
| Creators | Talone, Andrew |
| Publisher | University of Central Florida |
| Source Sets | University of Central Florida |
| Language | English |
| Detected Language | English |
| Type | text |
| Format | application/pdf |
| Source | Electronic Theses and Dissertations |
Page generated in 0.0016 seconds