Document Type

Other

Publication Date

10-2009

Abstract

Robot teleoperation interfaces are mostly graphics based nowadays, that is, all the information the operator receives is presented on a screen. Complex robot interfaces may lead to operator cognitive overload. The use of other senses to receive part of the data sensed and transmitted by the robot may help reduce this overload and thus enhance the performance of the operator. This paper aims at measuring the benefits of using vibrotactile feedback in Human-Robot Interaction (HRI) interface design specifically for an urban search and rescue (USAR) system. Our hypothesis is that the use of collisionproximity feedback interfaces (CPFs) should lead to an improvement of an HRI system performance by increasing the operator’s situation awareness (SA) and reducing cognitive load. Additionally, it should be flexible enough to be adapted to different HRI USAR systems. A user study encompassing a search task was performed to evaluate this new interface. An in-between subjects experiment tested the effect of both graphical and vibro-tactile CPF interfaces on performance in a simple search task in a virtual collapsedbuilding environment. Performance and situation awareness were measured based on task time, number of collisions with the environment, number of objects found and correct report of environment using sketchmaps. First, the results of this research highlight the importance of a homogeneity verification in the experiment groups, which is generally not reported in most research results but that can drastically affect the results of an experiment. Second, our results indicate how previous experience can affect subjects’ performance. Videogame experience seemed to have a slight impact on the performance of subjects. Third and most importantly, our results have indicated that the use of both vibrotactile and graphical feedback interfaces may improve operator’s performance in a search environment, and may indicate an increase in operator situation awareness (SA). Future enhancements in our system to better approximate it to a real robot control experience will help us consolidate the initial results obtained here.

DOI

WPI-CS-TR-09-09

Share

 
COinS