Skip to content

Latest commit

 

History

History
20 lines (16 loc) · 1.38 KB

README.md

File metadata and controls

20 lines (16 loc) · 1.38 KB

README

Code for generating fair trajectories for robots.

Scripts

  • experiment_central.py: runs a set of experiment trials for finding fair trajectories using a centralized method
  • experiment_distributed.py: runs a set of experiment trials for finding fair trajectories using a distributed method
  • objective.py: defines class responsible for formulating the optimization problem that finds fair trajectories
  • run_experiments.sh: runs a series of experiments for different notions of fairness and robot team size
  • generate_trajectories.py: generates the robot trajectories using code excerpt from https://github.com/markwmuller/RapidQuadrocopterTrajectories stored in trajectorygenerationcodeandreas

Fairness Notions

The scripts experiment_central.py and experiment_distributed.py take in a --notion argument for the fairness notion to be optimized over robot inputs u in the experiment. The following notions are implemented:

  • set --notion 1: u^TQu, an energy term is optimized where Q is positive definite.
  • set --notion 0: u^TQu + f1 is optimized, where f1 is the variance of the normalized energy of the inputs
  • set --notion 2: no energy term or fairness notion is optimized
  • set --notion 3: u^TQu + f4 is optimized, where f4 is the variance of the total energy surge of the inputs
  • set --notion 4: only f1 is optimized
  • set --notion 5: only f4 is optimized