search

CURG-archive / graspit_handop

Other
1 stars 2 forks source link

write code to analyze repeated planning sanity test #20

Closed jon-weisz closed 11 years ago

jon-weisz commented 11 years ago

for each hand, calculate the average and standard deviation of each of the grasp scores as we filter grasps by planning time

Essentially we want to be able to say for hand x, how different would the results have been from generation 0 to generation n if we hand only planned for t seconds.

What we want to be able to do is plot a mean difference and mean standard deviation of hand quality scores vs planning time

jon-weisz commented 11 years ago

First read this: http://www.engr.ucsb.edu/~shell/che210d/numpy.pdf

then read examine_database.py

then look at the bottom of eigenhand_db_tools.py and run_experiment.py:output_results

Together these should give you enough information to make a start of it