TODO.md

TODO List

This is a list of tasks that need to be done as well as possible enhancements

Single Robot Exploration

• Implement proper MDP algorithm
  • Add reward propogation into hex_grid
  • Define list of robot actions
  • Change Robot.choose_next_pose to choose based on rewards
• Allow for robots of varying sizes
• Asynchronously scan and update map while choosing actions in another thread
• Dynamically increase horizon if no reward within horizon

Multi-Robot Exploration

• Synchronously create multiple robots with perfect communication
  • Design map merging technique
  • Design basic collision avoidance
  • Update Markov Decision Process for all robots to operate synchronously
• Allow robots to operate asynchronously
• Modify MDP to allow for imperfect communication
  • Create Voronoi diagram of map (NO LONGER USED)
  • Modify MDP using distributed value functions for imperfect communication

Enhancements

• Add RangeFinder that works more realistically
• Make dynamically growing map i.e. no previous knowledge of map size
• Add negative reward to obstacles so robots stays away from them (by adding noise to movements in MDP)
• Add collision checking to world so robots don't accidentally enter objects
• Improve efficiency
• Add comprehensive unit testing
• Add comprehensive integration testing