After discussing some of the possible projects I could work on, I chose the
following. This is the first draft of my project proposal. I will continue to
add to this to develop my final paper. To see the updated version, click
1. Statement of the Problem
NUMACK is an embodied conversational agent, a virtual human who is capable
of interacting and communicating with people. He is a robot designed to give
directions around the Evanston campus of Northwestern University using speech
and gestures. In order to successfully give directions, NUMACK should also
incorporate a graphical representation of the campus. This map will draw and
animate the route while NUMACK is explaining it, allowing his directions to be
2. Previous Research
Based on the paper “Pictorial and Verbal Tools for Conveying Routes” by
Barbara Tversky and Paul Lee, we know that when given a tool kit of objects to
draw a map, people like to include landmarks along the way. This use of landmarks
shows they are just as important when explaining a route as the roads and turns
on the route.
The previous version of MACK used a projector to display a simple path on a paper
map. This new map would illustrate the route in real time and incorporate the
important landmarks. Thus, the representation of the route would be more visual
and interactive, leading to a vast improvement on the previous version of NUMACK.
3. My Approach
I plan to use Open Inventor to create a graphical representation of the
Northwestern campus. After learning how to use Open Inventor, Amanda and I will
look at some of the video data to determine a few routes, including landmarks.
We will use this as a starting point for both of our projects.
I will then begin by creating a simple map using these routes and landmarks.
Once the map is finished, I will focus on taking one of these routes as input and
drawing it on the map. This route will be determined from Amanda’s work, so we
will need to discuss the format of this input/output. The drawn route will consist
of animations such as highlighting, drawing arrows, and circling important landmarks
along the route while NUMACK is talking about them.
As we are implementing our separate projects, Amanda and I will also be thinking
about how people choose which landmarks to mention as they describe a route. Since
we need to decide on landmarks for the rest of campus, the information we gathered
from the video data will be helpful.
Once the map is working for a small set of routes, the final step will be to
expand the map to include these landmarks across the remainder of the campus.
4. End Goal
I believe this project will be complete when I have a good representation of
campus that is helpful in expressing the correct route from point A to point B. An
important part of this project is determining what is significant and should be
included in the map when drawing a route from a given set of directions. Also, the
animations pointing out the route and landmarks is a key part of the map. These two
points will be good measures of a successful project.