CogEnvDecoder 1.0.3

For decoding observation data from Cog virtual environment

Background

This is a package for COG participants to get necessary image and vector data to train and test their models. ##Install pip install CogEnvDecoder ##Usages The code is able to get observations generated from ml-agents in the COG stimulation environment. With the code, you can get:

• Blue one rover
  • Image generated from virtual camera setted up on Blue one
  • Blue one rover position & angle
  • Enemy rover position & angle
  • Remaining HP & bullets
  • Goals' position
• Judgement system information
  • Current score
  • Number of goals have been achieved
  • Whether enemy rover has been activated
  • Enemy rover remaining HP & bullets
  • Blue one caused damage
  • Time passed in the round
  • collision information