Design a Fitness Score to grade Filter Performance

DavidSmith166 / mrover-workspace

The University of Michigan Mars Rover Team workspace.

2 stars 0 forks source link

Design a Fitness Score to grade Filter Performance #15

Open DavidSmith166 opened 4 years ago

DavidSmith166 commented 4 years ago

Work by yourself and determine how would best score the performance of our filter considering out mission parameters. The input to your algorithm should be an array of GPS locations for the truth, and the filter output. Bonus points if it is fast.

warussel commented 4 years ago

Most significant issue that I have run into is that the truth data and the output filter data don't have a perfect one to one matching, which makes it difficult to compare the two data sets from a statistical standpoint. Currently looking into using numpy's griddata function to interpolate the truth data into a function that we can index into with our filtered data and then compare the resulting values.