Analysis

The results from the Robotics class were not what we expected. A fair amount of good feedback was received, but a lot of extraneous information was received from participants. A significant amount of the information received did not apply to our project. For instance, we had a comment that said, “It has to have the chair in certain position. It’s better to find a chair leg autonomously.” First of all, it does work autonomously; we have no interference with it once the program is running. We had to test the chair in the direct approach position because, for some reason, it did not work when the chair was positioned to where the robot has to maneuver around it. It worked in prior tests, but not during the presentation. We received a few comments that were along the same lines and we chose to discard them as questionable feedback. Another comment of this type dealt with a problem with the camera. For some reason, the camera did not want to tilt when we told it to, so we had to manually tilt it. We tilted it a little farther than it was in our code, so it saw part of the laser range finder. The problem with this is that the laser has little arrows on the top of it that just so happen to be the same color as the leg of our chair (a manila color). We got a comment about this, because the robot recognized those colors on itself as a leg of the chair.

Another type of comment that was received dealt with other methods to recognize a chair. We had some people say we should have used “other distinguishable properties instead of color,” along with using edge detection. We considered these “other distinguishable properties,” but with time constraints, color was the best choice for our project. Using edge detection, or some sort of algorithm such as this, would be a much more complicated task.

We did get some positive feedback as well. Someone expressed the opinion that the chair gripping module appeared to be the hardest to implement. This individual thought we did a good job with this aspect of the project. Another comment dealt with the ACTS system, saying that it seemed to work well with our project. One person even understood the concept of our project a little better than most. This individual said something to the effect of wanting to see our project once we eliminated bugs, because it would be interesting.