-
Notifications
You must be signed in to change notification settings - Fork 14
Workshop 7 ‐ Comparative statistics
First, synchronise your fork with the main repository on GitHub (Sync fork) as some structural changes were made and new files were added. Then clone or pull the repository to your local PC and re-open it in the dev container using VSC. Rebuild the updated packages colcon build --symlink-install, and source the repository source install/setup.bash.
The evaluation folder contains a comp_stats.py file illustrating the use of different statistics covered in the mini-lecture. The script features two data sets A, B with numbers generated from a normal distribution (normally these would be your experiential results, e.g. object counts from individual experiments). Inspect the file and check how the key metrics are being implemented including standard error, one- and two-sample t-tests and how to obtain p-values.
-
Interpret the results using the p-values and decide if the results of the tests confirm or reject the alignment (null hypothesis). Discuss the results with your peers.
-
Experiment with different values of N and N_short (e.g. 40 and 10, 100 and 20) to see how the results and metrics change paying attention especially to p and t values.
-
Consider, how to use these tests for the assignment.
This task is designed to help you create custom simulated scenarios needed for the assessment. The rob2002_tutorial package contains now a folder called worlds which contains a simple custom-made Gazebo world called custom_world.world which you can use it with the simulated limo robot by running the following command: ros2 launch limo_gazebosim limo_gazebo_diff.launch.py world:=src/rob2002_tutorial/worlds/custom_world.world.
You can easily create your own worlds in the Gazebo simulator (to run it, simply type gazebo in the command line) starting from scratch. To test that functionality, add a number of objects to the scene (see Workshop 2 for some examples) and then save the world in the src/rob2002_tutorial/worlds folder. Relaunch the robot simulator with your newly created world to test that everything works as expected.