Real-Time Human-Robot Interaction Issues
The virtual 3D display using an ego centric perspective.
The inclusion of human-robotic systems into operational scenarios brings about many issues in Human-Robot Interaction (HRI). These issues vary depending on whether the robot is operating in proximity of a human operator or if the robot is being monitored and 3D-Display controlled from a remote location. In general, these issues include situation awareness, vigilance in monitoring, mode confusion in control, and trust. Additionally, past interface and control systems for remote robotic systems have been complex in nature. In order to reduce the amount of training required to operate such robotic systems, it is necessary to enable more natural interactions between humans and robots. This project will use multi-modal interfaces that include speech and natural language understanding, audible warnings and vibro-tactile alerts in addition to graphical and textual components. A new virtual 3D display allows the operator to visualize the robot within a world representation that emerges as the robot moves.
This remote operator console will be placed within a mobile van that can drive on location. For many tasks, however, a large console is not necessary or possible. Operators who need a more portable interface can use a more compact interface that combines the various interface screens into one 1024 x 768 visual display.
Currently, the graphical interface shown above is used on a light-weight, portable touch screen. Special attention has been paid to how novel man-machine interfaces can explicitly reduce uncertainty for both human and robot. Humans should not have to guess at the intentions of a robot, and conversely, robots should not have to disambiguate or make inferences regarding uncertain human input. Interfaces that accomplish these goals will encourage operator trust and reduce human / robot error.