Codebase for the paper titled "Active Localization of Multiple Targets from Noisy Relative Measurements" (WAFR 2020)
Check out our project page!
For training a session, run:
python -m target_localization.train --sess atl --num_targets 2
Add the flag --image_representation
to use the encoding of the the belief-map image as part of the observation function. Refer to the paper for more details.
For evaluating the agent policy, run:
python -m target_localization.eval --sess atl --num_targets 2