Active Scene Recognition folder consists of different packages, including both the functionalities used by it as well as its logic (in a state machine). Its main goal is to recognize scene(s) in distributed and cluttered search spaces on which scene recognition alone would fail. Our definition of scenes consists of objects which are connected to each other by means of spatial relations.

• asr_direct_search_manager: Handles the views for the direct-search mode of Active Scene Recognition. The direct search can be executed without prior knowledge.

• asr_grid_creator: Generates a grid for the asr_direct_search_manager to determine the views which are to be used in the direct search mode.

• asr_intermediate_object_generator: Weights the different objects by their relations to other objects. The weights can be used in the asr_next_best_view to prioritize objects above others or to search only for a subset of objects (intermediate objects) in the direct search.

• asr_lib_pose_prediction_ism: This package contains classes and algorithms to predict poses of searched objects by the help of a tree of ISMs.

• asr_next_best_view: This package estimates Next-Best-Views as well as configurations (target positions and orientations) for a robot.

• asr_recognizer_prediction_ism: This package includes the algorithmic link between object pose prediction and scene recognition, both based on ISM trees.

• asr_recognizer_prediction_psm: Contains a combination of pose prediction and scene recognition for the alternate scene representation PSM.

• asr_resources_for_active_scene_recognition: This package contains various helpers and resources for the active scene-recognition applications. This package is one of the two entry points for active scene recognition.

• asr_state_machine: This package implements the logic which connects different functionalities such as (passive) scene recognition, object pose prediction and Next-Best-View estimation to Active Scene Recognition. This package is one of the two entry points for active scene recognition.

• asr_world_model: This package implements the logic of our Online World Model, while the Offline Knowledge Base is distributed across asr_object_database, mild_navigation and asr_visualization_server.

• .