Touchscreen Usability in Microgravity
Jurine A. Adolf*, Kritina L. Holden **
- *Lockheed Martin Engineering and Sciences Services
- 2400 NASA Road 1, C81
- Houston, TX 77058-3799
- +1 713 483 9724
- **Syntropy Corporation
- P.O. Box 891526
- 15407 Lanswick Dr.
- Houston, TX 77062
- +1 713 286 2628
- Human Factors and Ergonomics Laboratory
- Frances E. Mount, Manager
- NASA Johnson Space Center
Touchscreen technology is well-suited for extreme environments, for example, microgravity. However, the usability of touchscreens has not been tested in this environment. The Human Factors and Ergonomics Laboratory (HFEL) at the NASA Johnson Space Center has conducted three evaluations of touchscreen usability both in a simulated weightless environment and on a space shuttle mission. Preliminary findings suggest that touchscreens were preferred for those tasks with larger touch areas, but not for precise positioning. Not anticipated though was the hand fatigue experienced by astronauts. Complete results will be available.
Touchscreen, Input Devices, Cursor Control Devices
Touchscreen technology is currently being used successfully in one-gravity (1-g) environments for tasks requiring limited inputs, such as information kiosks, cash registers, and process control applications [1,2]. Because touchscreens are direct manipulation devices, they are faster, easier to learn and more flexible. Additionally, they have no moving parts and require no extra workspace [2,3]. This makes touchscreens especially well-suited to extreme environments, for example, submarines, aircraft and space. However, the usability of touchscreens has not been tested in all of these environments, especially microgravity. Advances in hardware and software have solved many of the problems (e.g., glare, parallax, limited tactile feedback, resolution, stabilization) associated with touchscreens in 1-g environments , but the solutions may not be adequate for all environments. The Human Factors and Ergonomics Laboratory (HFEL) at the NASA Johnson Space Center has tested the usability of touchscreens to identify problems in microgravity (e.g., need for restraints) and to develop display guidelines (e.g., minimum object sizes).
Hardware and Software Description
An Elo TouchSystems AccuTouch(r) (model E274) resistive membrane touchscreen integrated with a 9.4" active matrix color TFT liquid crystal display (LCD) monitor (model LMT 5020) was used in the investigations. The touchscreen has more than 100,000 touchpoints per square inch and requires an activation force of 3 to 4 ounces. Resolution of the monitor is 640 x 480. The monitor was connected to the external VGA port of an IBM Thinkpad 755C. Experimental software was written in Asymetrix(tm) Multimedia Toolbook 3.0.
An informal evaluation and follow-up experiment of touchscreen technology were done in a simulated weightless environment (KC-135 aircraft). The KC-135 follows a parabolic flight path which produces brief (30 sec) periods of weightlessness at the top of the arc.
Phase I Approach
The first evaluation was completed to test the functionality of the hardware in the KC-135 environment and to investigate the need for hand and/or foot restraints. Seven subjects used the touchscreen to perform common computer tasks with existing Microsoft(r) Windows applications, such as, Paintbrush, Solitaire, calculator, and Word for Windows(tm).
Phase I Results
No hardware problems due to the KC-135's simulated microgravity occurred. Most subjects felt that a foot restraint alone was sufficient for most of the tasks. However, when constant touch pressure was required for tasks, such as drawing or dragging, both hand and foot restraints were necessary. Some problems were experienced with the hand restraint due to its location on the side of the touchscreen.
Subjects preferred the touchscreen for those tasks with larger touch areas, for example the calculator and solitaire applications. For precise positioning, such as placing the cursor between two letters, subjects found the touchscreen more difficult to use.
Phase II Approach
The second KC-135 evaluation compared the usability of the touchscreen with a Toshiba pen-based system in both the ground and microgravity environments. The pen-based system was included in order to compare devices for another NASA project. For each device, the six subjects completed the following sequence of tasks: two pages of single click pointing, two pages of double click pointing, and three pages of text editing. Variables manipulated in the clicking tasks included size of active touch area (i.e., button size) and inter-object distance. Sixteen levels of button size in both square and rectangle shapes were tested. The buttons were separated by either a 1/4" or 1/8" interdistance. The text editing task required subjects: (1) to select a word by double clicking the text, (2) to select a word or phrase by drag/selecting the text, and (3) to type and replace selected text. Only a 12-point font size was tested. These tasks were chosen as representative of most basic computer input tasks. Subjects were encouraged to use hand and/or foot restraints as necessary.
Dependent measures included task time, errors, and subjective comments. Errors were further analyzed to determine if direction of movement caused overshooting or undershooting of the target. Also, for the double click pointing task, timing errors (e.g., double click interval too long) were counted.
Phase II Preliminary Results
Descriptive statistics suggest that subjects performing the tasks in the KC-135 environment took longer and made more errors than on the ground. In comparing the devices, subjects made more errors with the touchscreen. However, it appears that the larger number of touchscreen errors occurred during the KC-135 portion of the study. There does not appear to be much difference between devices in terms of reaction time. The smallest two button sizes (the smaller about the size of a scrollbar arrow) seem to account for the most difficulty in clicking. A frequency count of the type of errors indicates that subjects most often missed the target by touching directly above it.
Even though subjects were trained prior to the experiment, a repeated Analysis of Variance (ANOVA) showed a significant effect over blocks of trials for both reaction time and errors. Therefore, further inferential statistics will only use those blocks where a stable level of performance was achieved. Those analyses will be completed for the poster presentation.
SPACE SHUTTLE EVALUATION
An inflight evaluation by Space Shuttle (STS-70) astronauts compared the touchscreen to the standard onboard input device, the IBM Thinkpad Trackpoint II(tm). All five crewmembers performed two ground baseline and two onboard data collection sessions for each device. Each session consisted of the following order of tasks: single click pointing, double clicking pointing, dragging, and text editing. For the pointing and dragging tasks, five button sizes corresponding to standard Microsoft Windows objects were tested. The text editing task required the subjects: (1) to select a word or phrase by either double clicking or drag/selecting, (2) to select a menu item (e.g., bold), and (3) to type and replace selected text. Three font sizes (10, 12, 14) were tested.
The same dependent measures collected in the Phase II KC-135 study were recorded. Following each session, the crewmember responded to an electronic questionnaire. Additionally, the onboard sessions were videotaped.
Astronauts were debriefed following the completion of the flight. Some of the crew preferred the touchscreen on the ground, but the Trackpoint in zero-gravity. Many experienced hand fatigue almost immediately only in the zero-gravity environment. Descriptive and inferential statistics will be completed for the poster presentation.
One of the disadvantages of touchscreens in 1-g environments is arm fatigue [1,3]. In the zero-gravity environment, this fatigue should not occur. Surprisingly, the astronauts did experience hand fatigue, possibly due to the nature of the experimental setting. For example, during the several trials of clicking, double clicking and dragging, a subject might not tend to rest or reposition the hand like they would most likely do when performing a higher-level task (e.g., word processing, spreadsheet). A follow-up study of this issue is planned.
Overall, the touchscreen performed well for the larger target areas. However, its performance in microgravity was not equal to that on the ground. Following completion of the descriptive and inferential statistics for both the second KC-135 and shuttle evaluations, design guidelines on minimum touch areas will be developed for this type of environment.
- Shneiderman, B. Touch screens now offer compelling uses. IEEE Software, 8, 2, (March 1991) 93-94, 107.
- Potter, R., Weldon, L. and Shneiderman, B. Improving the accuracy of touch screen: An experimental evaluation of three strategies. Proc. CHI'88. (Washington, D.C., May, 1988), ACM Press, 27-32.
- Sears, A., Plaisant, C., and Shneiderman, B. A new era for high precision touchscreens. In Advances in Human-Computer Interaction, 3 (1992), Hartson, R and Hix, D. (Eds.), Ablex, NJ, 1-33.
- Sears, A. and Shneiderman, B. High precision touchscreen: Design strategies and comparison with a mouse. Int. J. of Man-Machine Studies, 34 (1991), 593-613.
 Work completed during employment at LMES