Scalability of Robotic Controllers: Effects of Progressive Levels of Autonomy on Robotic Reconnaissance Tasks

Report No. ARL-TR-5258
Authors: Rodger A. Pettitt, Elizabeth S. Redden, Estrellina Pacis, and Christian B. Carstens
Date/Pages: August 2010; 70 pages
Abstract: This investigation of the effects of progressive levels of autonomy on robotic reconnaissance task performance was a cooperative effort between the U.S. Army Research Laboratory/Human Research and Engineering Directorate (ARL/HRED) and Space and Naval Warfare Systems Center San Diego (SSC Pacific). Thirty Soldiers served as participants. After training on the operation of the robotic system, each Soldier completed building reconnaissance exercises using three different levels of robotic automation (teleoperation, semi-autonomous, and autonomous operation). The participants' primary tasks were to map the floor plan of the building, and to identify and photograph objects of interest (situational awareness tasks). In order to increase the cognitive demand, Soldiers responded to requests for information while operating the robot. Dependent variables included objective performance data, data collector observations, and Soldier questionnaires. Reconnaissance times were significantly faster and there were significantly fewer driving errors when the robot was in the autonomous mode than when it was in the other two modes. Performance in the semi-autonomous mode was significantly better than in the teleoperation mode. As the level of autonomy increased, workload, reconnaissance times, and driving errors decreased and mapping accuracy increased. There were no significant differences among the conditions in situational awareness or target identification.
Distribution: Approved for public release
  Download Report ( 1.144 MBytes )
If you are visually impaired or need a physical copy of this report, please visit and contact DTIC.
 

Last Update / Reviewed: August 1, 2010