After last week's failed testing, I wasn't sure where to go next. The first thing I did was to implement the plan to make new text files when control of the cursor switched from the mouse to the Camera Mouse. This involved going into the code and looking up documentation of C#. After doing this, I thought back to the testing session. One of the requests for my original test that the caretaker of the students had suggested was to make it easier by making fewer targets and making the area required for the window smaller. This way, it would be easier for the students to access each target. I did this (see here) each button is pressed. Hopefully when we return to the Campus school in a few weeks, the students will be able to successfully navigate this simpler test program.
This week I also received a program from a man working here at BU (learn more about him here) with my data, the two paths were the coordinates of the mouse at each frame, and the coordinates where the mouse was clicked. This means that the paths I want to compare are the shortest straight-line distance between two mouse clicks and all screen coordinates between these two clicks. This means in any given data set, there are 700+ points in the set of screen coordinates, but only 14 points in the set of mouse click coordinates. This means that the DTW distance would be calculated like this:
Meanwhile, what I wanted to calculate would look like this:
So I wrote my own program to do this. First, for each line between two vertices, I determined if the line he line was horizontal, I looked at the x coordinates of the screen set. If the coordinates were between the x coordinates of the two vertices, I found the point on the line that was perpendicular to the straight-line path that also went through the screen coordinate and calculated the distance between the two points. If the coordinates were outside the boundaries of the vertices, I calculated the distance between the coordinate and the nearer of the two vertices. If the line was vertical, I did the same thing, but with the y coordinates. I added up all these distances to get the total distance from the shortest path, and also divided by the number of points to get the average distance from the shortest path.
Now, to use this program along with the DTW program and my length comparison program, I decided to compare trials of my movement analysis test using the mouse and the camera mouse. To do this, added code to track the mouse movement when the cursor was controlled by the mouse and not the camera mouse after pressing the Caps Lock key.