Alle Publikationen
- <<
- <
- 1
2018
-
(2018) : Expression of intention by rotational head movements for teleoperated mobile robot: 2018 IEEE 15th International Workshop on Advanced Motion Control (AMC): 2018 15th International Workshop on Advanced Motion Control (AMC): Tokyo: 3/9/2018 - 3/11/2018: [S.l.]: IEEE, S. 249-254
Abstract: We are studying a teleoperated mobile robot that provides useful information to a pedestrian. However, it is difficult for people to understand meanings of actions, motions or movements of many conventional robots. The purpose of this study is to improve pedestrian's impressions of a robot. Especially this paper describes people's understandability of robot behaviors when a robot turns around a corner or when a person and a robot pass each other in a corridor. Our robot shows its intention to make turn by rotating its head, as though a pedestrian shows a traveling direction by his/her gaze or face direction. The robot is teleoperated by an operator for safety in public spaces, and the direction of the robot head and the moving direction of the robot body are determined by an artificial potential field (APF) generated by a target position given by the operator, positions of obstacles and pedestrians. The APF for a pedestrian is generated based on her/his personal space of a person. Thus, the robot can express the intention of its action by rotating the head to look where it is going, when the robot changes its direction around pedestrians. The intention expression can be natural and understandable for them by the rotational movement of the head before the robot turns its body actually. Impression evaluation experiments with questionnaires were conducted under the two kinds of situations to reveal the validity and effectiveness of the intention expression by the robot's head rotation. Significant differences related to understandability and some impression words were observed between with and without rotating the head.
2013
-
(2013) : Feature-Based Prediction of Trajectories for Socially Compliant Navigation In: Roy, Nicholas; Newman, Paul; Srinivasa, Siddhartha (Hg.): Robotics: Science and systems VIII: Cambridge, Massachusetts: The MIT Press
DOI: https://doi.org/10.7551/mitpress/9816.003.0030 Abstract: Mobile robots that operate in a shared environmentwith humans need the ability to predict the movements ofpeople to better plan their navigation actions. In this paper, wepresent a novel approach to predict the movements of pedestrians.Our method reasons about entire trajectories that arise frominteractions between people in navigation tasks. It applies amaximum entropy learning method based on features that capturerelevant aspects of the trajectories to determine the probabilitydistribution that underlies human navigation behavior. Hence, ourapproach can be used by mobile robots to predict forthcominginteractions with pedestrians and thus react in a socially compliantway. In extensive experiments, we evaluate the capability andaccuracy of our approach and demonstrate that our algorithmoutperforms the popular social forces method, a state-of-the-artapproach. Furthermore, we show how our algorithm can be usedfor autonomous robot navigation using a real robot.
- <<
- <
- 1