Track Variance

In lab there are four different tracks across the room that AEV’s will run on, because all of these tracks are different the AEV will perform differently. Using one code on the same AEV for all four tracks the team will:
  1. Collect data from five runs on each of the four test tracks.
  2. Calculate statistics based on the variations between the test tracks.
  3. Draw conclusions based on the calculated statistics about how similar the different test tracks are.

The single code used for all 20 runs can be found under the “Arduino Codes” tab. The code was set up to run the motors up the inclines of all the rails, and then coast to the gates. This separation in the code provided an opportunity to further dissect the data into two columns based on the motors running or coasting.

Table 3 above shows the distance, time, and energy for all 20 runs at the moment when the motors were shut off, and at the very end once the AEV was completely stationary.

Table 4 looks into more advanced statistics for the runs such as mean, median, range, and standard deviation.

Conclusions Drawn:

  • From Table 3 & 4 above the information is inconclusive in terms of significant differences. By looking at the data once the motors stopped one can see that the distance was almost the exact same every single run with very little difference in time or energy between tracks. After coasting there is a slightly larger spread in data, but that spread on each individual track seemed to be the same no matter which track the AEV ran on. Because of this the team could not conclude and significant difference between comparing the tracks throughout lab.

Assumptions Made for Testing:

  1. The dependent variable (time) is measured on a continuous interval.
  2. The independent variable consists of two categorical independent groups (distance & energy).
  3. There is independence of observations with no relationship between the observations of the groups. Meaning, there are no participants in more than one group.
  4. There should be no significant outliers.
  5. Your dependent variable should be approximately normally distributed for each category of the independent variable.
  6. There needs to be homogeneity of variances (the variance within the populations is equal).