Design Solution
Overview
Our final solution was informed by various sources of data we obtained in our iterative process of design. These data sources included the simulation requirements our clients generated from our user research as well as HCI methods that uncovered usability issues.
Multiple Views
The simulation our users must complete requires them to access different types of information at various points in time. They need to plan the rover's itinerary, decide which targets to capture, access images taken, correct for rover error, and keep track of the time passed. To this end, we customized different components and views to address the needs of the different tasks our users have to do. Our software enables the scientist and engineer to use separate physical computers and monitors, since the tasks of correcting for rover error and analyzing images are performed in parallel. A third physical display shows the participants' individual and combined scores.
Engineer's Activity Planning View
Overhead Map: Allows the user to visually author a series of commands for the rover
Command Sequence View: Shows a detailed view of the rover's sequence of commands.
Image View: Shows the most recently updated picture.
Timer: Shows the simulation time and when downlinks will occur.
Scientist's Scoring Map & Image View
Overhead Map: Map shows adjacent sides and allows the user to annotate score as it updates on the scoreboard
Image Library: Updates after every downlink
Correcting for Rover Error
Before every downlink, the rover allocates enough time to take a pancam shot, a 360 degree view of the rover's surroundings stitched up from multiple pictures. Our initial user studies (during CW) indicated that users had the most success dealing with a pancam that was circularly warped since this view most directly relates to the rover's position.
Upon receiving a downlink, the user goes into the error correction mode which where the user can perform triangulation by specifying three targets from the pancam just taken and matching them with the same three targets on the map. Users can rotate the pancam in case the predicted orientation is off. Targets are matched by clicking them with the right modifier keys.
Engineer's Error Correction View
Direct Manipulation
Users are interested in higher-level goals, such as taking pictures of boxes, not minute details such as what the camera's tilt degree should be. To support these high level tasks, our design abstracts away many of these low level commands.
Creating a drive plan
Users can create an itinerary for the rover by using the drive tool we provide. A cursor allows the user to preview the next set of commands the user will author before committing to them by clicking down. Users can add points in the middle of drive paths by pressing the keyboard path button to toggle the mode. The user can easily delete points by holding down a modifier and clicking the point. The user can easily drag the location of this icon to command the rover to drive to a specified location, automatically updated the previous commands to turn and drive the rover the appropriate distance.
Taking pictures of Targets
The main goal of the user in this simulation is to take pictures of targets out in a field. As a result, we provide a picture tool that allows the user to easily drive up to and take pictures of targets. The head of the rover can be dragged to change the direction the rover to face the target. The user clicks and drags the yellow sector to specify the camera's pan and tilt in terms of the field of view area the user expects to capture.
Situational Awareness
The simulation our users must complete requires processing lots of information in a limited time. Because of this high cognitive load in such a stressful environment, we designed our interface to allow the user to easily access, manipulate, and be notified of information.
Receiving notification of downlinks
The progress bar at the bottom of the interface indicates how much time has gone by. The normal color is green; when a downlink is about to occur, the timer will change to red. In addition, an alert sounds right before downlink to notify the user that a downlink will happen.
Receiving notification of rover status
The commands in both the map and sequence view are color coded to indicate the status of the commands: white indicates not yet uplinked, orange indicates uplinked not yet completed, and gray indicates completed. The only commands that are editable are commands not yet uplinked.