Miller Robotic Interface II Manuel d'utilisateur

In Drury et al. [2007], human-robot awareness is further broken up into five types to aid in assessing the operator’s SA. The categories are location

the operator accidentally driving the robot out of the arena into the crowd and bumping into a wall trying to get back into the arena. The turned came

1.2 Contributions This work describes an interface design that is a result of an evolutionary process. This design was validated through user testin

to make real-time decisions should be presented on, or around the main video screen, so the operator does not have to look far to see it. Prior to

2 REVIEW OF RELEVANT RESEARCH Remote robot interfaces can be broken down in to two categories: map-centric and video-centric. A map-centric interfa

Figure 1: MITRE’s Map-centric Interface. Photo from: [Yanco and Drury, 2007] The Idaho National Laboratory (INL) also has a map-centric interface,

the actual area looks like. Also, it can be hard to get good surroundings awareness due to how the walls are drawn with respect to the robot avatar.

2.2 Video-centric Interfaces Video-centric interfaces are by far the most common type of interface used with remote robots. It has been shown in st

Figure 3: ARGOS interface from the University of Brno. Photo from: http://www.orpheus-project.cz/galery/img/5/5_slideshow.jpg The interface designe

Figure 4: Swarthmore College’s interface inspired by first-person shooter video games. Photo from: [Keyes et al, 2006] Figure 5 Figure 5: On

EVOLUTION OF A TELEPRESENCE ROBOT INTERFACE BY BRENDEN KEYES ABSTRACT OF A DISSERTATION SUBMITTED TO THE FACULTY OF THE DEPARTMENT OF COMPUTER SCIENC

Commercial robot interfaces such as iRobot’s Packbot, Foster-Miller’s Talon, Remotec’s Andros and American Standard Robotics’ VGTV Extreme, two of wh

3 TESTING METHODOLOGY Because it is important to quantify situation awareness (SA), we discuss SA measurement techniques here. Again, situation awa

in these studies. The post run questions do ask how they felt they did in completing the task. These questions are subjective forms of SA. However,

procedure often used a script to be sure that the mission was presented identically to each test subject. We wanted to simulate the stress of a real

questions to the subject, if they were being too quiet or seemed confused, so the administrator could pinpoint in the notes what was the cause of the

asked to perform this secondary task. The ability to traverse their way back directly showed if they had good location awareness. In Chapter 5, we

4 THE ROBOT SYTEM Although this thesis describes the evolution of the robot system as a whole, the majority of changes to the system, as a result of

We noted during runs of other systems we were studying that a large percentage of the robot hits in the environment were directly behind the robot [Ya

camera is used to detect heat signatures and help find victims that may be covered in dust as well as in darkened areas. We did some work trying to f

Figure 7: Interface control architecture. The video uses the Phission package to capture the video on the robot end. Phission is a concurrent visi

ACKNOWLEDGEMENTS I'd like to thank everyone who has supported me over the past few years. First and foremost, this is dedicated to my mother, wh

send the corresponding command out to the robot to be acted upon. These events control many actions, including moving the robot, changing autonomy mo

The system does have the ability to be expanded to use more autonomy modes, as shown in Casey, Chanler, et al. [2005]. However, adding more autonomy

Figure 8: Older ARGOS interface from the University of Brno. It uses center crossed lines to indicate the pan/tilt status of the camera. Photo fro

robots bumped obstacles in the environment an average of 2.6 times per run. Also, of the 29 total hits that occurred in the study, 12 or 41%, of the h

Figure 9: The map panel. It shows the current area that the robot is in, as well as the path it took around the environment. The robot is indicate

Figure 10: The mode panel shows the current autonomy mode the robot is in. The background color and highlighted button will change depending what

promote proper interface use. Recall, this was a guideline as proposed by Yanco et al. [2004]. These suggestions ranged in variety. Some suggestion

5 INTERFACE EVOLUTION This chapter describes the interface evolution. It discusses how the interface is laid out and why. It also tells about the

The original prototype consisted of many of the panels described in section 4.4. The large video screen can be seen on the left center of the screen.

hypothesized that having both the color changing, as well as how many bars were filled in, would allow for a better understanding of the robot’s situa

Portions of this thesis have been previously published in the following works: Brenden Keyes, Robert Casey, Holly A. Yanco, Bruce A. Maxwell and Yav

INL robot. The biggest advantage over the INL system, however, came from the added rear camera. There was another difference between the studies tha

from the front camera to the rear camera. This switch also remaps the driving controls and sensor information so that the back of the robot appears t

the robot’s path is blocked would be a possible solution to this irritation. We discuss this further in section 5.2 which discusses version 2.0. As

accept any suggestions said they were still helpful because they showed multiple interface features and when their use was appropriate. Another subje

environment. The bird’s eye view provided by the map can assist in producing a good mental model of the environment which leads to good location awar

In our design, we sought to reduce the number of collisions by surrounding the video screen with proximity information. We felt that having the impor

to the front camera’s video stream. Interface B consisted of all the same panels as interface A, but the user could switch the main video panel to di

For this study, we had 19 subjects ranging in age from 18 to 50, with 11 men and 8 women. Each subject operated the robot through the arenas three ti

We found that most of the hits occurred on the robot’s tires. Of all the front hits that occurred with the system, 75% of the time it hit with the tir

Although this study was not a study of the complete system since the suggestion system and map were removed, it did give us good insight to the parts

TABLE OF CONTENTS ABSTRACT...

better. Due to how much distance values change as the robot moves, the icons were constantly changing colors and the number of bars filled. The rapi

Figure 17: Interface screenshot showing the rotated distance panel caused by the user panning the camera to the left. The red boxes line up with th

previous version. With the 2D view, rotating the panel clockwise or counter-clockwise was programmatically difficult. However, with the 3D panel whi

INL interface. We wanted to test the various thoughts that a map-centric interface would cover more area than a video-centric interface. We also wan

One possible confounding variable for this difference was the size of the two robots. The ATRV-Mini (INL’s robot) was smaller than the ATRV-Junior (UM

The number of victims found was also compared between the two interfaces. We had hypothesized that the emphasis on the video window and other sensor

positive comments for INL identified the ability to have both a three dimensional and two-dimensional map. Subjects also liked the waypoint marking ca

comments (two for the sonar ring display blocks not being definitive and one for the FLIR camera). Our analysis suggests that there are a few categor

Figure 19: Top mini-map shown in the top-right of the screen. Bottom: Toggled large view of the map, showing a larger area. Screenshots taken fro

idea of how close an obstacle is to the robot without overwhelming the user. It is also extremely accurate, which can help produce a better mental mo

LIST OF FIGURES FIGURE 1: MITRE’S MAP-CENTRIC INTERFACE ...

Figure 20: Version 3.0 of the interface. It boasts a larger video window, a new distance panel and a relocated mode bar. Figure 21: The new dis

5.3.2 Experiment and Results We again performed an experiment on this version of the interface. This new study consisted of 18 users, 12 men and 6 w

We were studying the effects of the distance panel. If we allowed the robot to take some initiative, such as stopping itself, it may have prevented m

Table 3: Time and hit results from the study performed on version 3.0 of the interface Interface Time (s) σt Collisions σc A 507.9 283.6 8.8 3.7 B 63

When comparing the number of collisions that occurred, our initial hypotheses were correct. Interface A had the most collisions, with an average of 8

As for the new distance panel, the majority of the users (11 of 18) preferred interface C, while six of the eighteen participants preferred interface

6 RESULTS AND CONCLUSIONS We have succeeded through design and testing in providing a very useful surroundings awareness panel that displays accurat

We also add our own guidelines to enhance the list of proven guidelines. • Important information should be presented on or very close to the video sc

Table 4: Original Guidelines and Results Original Guideline Result Discussion Revised or New Guideline (if applicable) Provide a map of where the ro

7 FUTURE WORK The first and most important feature that should be implemented in the future is a mini-map. We have heard many complaints about the

LIST OF TABLES TABLE 1: SITUATION AWARENESS AND PERFORMANCE RESULTS...33 TABLE

displayed similarly to how the current distance panel is now displayed. This would make toggling from a perspective view to a top-down view much more

8 REFERENCES 1) Baker, M., R. Casey, B. Keyes and H. A. Yanco, “Improved Interfaces for Human-Robot Interaction in Urban Search and Rescue,” Procee

11) Endsley, M. R., “Design and Evaluation for Situation Awareness Enhancement,” Proceedings of Human Factors Society 32nd Annual Meeting, Santa Monic

23) Nielsen, C. W. and M. A. Goodrich. “Comparing the usefulness of video and map information in navigation tasks,” Proceedings of the Human Robot Int

33) Yanco, H. A., M. Baker, R. Casey, B. Keyes, P. Thoren, J. L. Drury, D. Few, C. W. Nielsen and D. Bruemmer, “Analysis of Human-Robot Interaction fo

APPENDIX A The design of the USAR robot system was a group effort: many people put many hours into various aspects of its design. This section is mea

A.1.2 Our Interface Studies There were four studies that were conducted on the interface versions presented in this thesis work. The first two and th

A.2 Robot Hardware and Software The ATRV-JR robot went through many changes during the course of this research. Michael Baker andrew Chanler and Phil

time Transfer Protocol (RTP). Robert Casey and I implemented the JMF on the robot. Philip Thoren installed his project, called Phission, on the robo

A.3 The Interface The interface is where I did the majority of my work. Just about everything on the interface was created and added by me. However,

1 STATEMENT OF THE PROBLEM 1.1 The Need for a Telepresence Robot Interface Robot operations are progressively becoming more important in a variety

icon for the suggestion panel was found on the internet at http://forums.macmerc.com/phpBB2/images/avatars/gallery/ikonboard/IconFactory5_robot.gif.

perception of elements in the environment within a volume of time and space, the comprehension of their meaning and the projection of their status in