Semantic World

Contents

Semantic World#

Welcome to the semantic world package! The semantic world is a python package that unifies the access and manipulation of scene graphs. It features a:

  • Unified API to scene graphs

  • Forward simulation for worlds and agents

  • Semantic interpretation of structures in the world

  • Automatic serialization of worlds to databases

  • Connecting worlds to different simulator backends through Multiverse [GN24]

This package originates from different developments of the AICOR Institute for Artificial Intelligence. Four different projects developed a very similar component for different parts of cognitive modules. This project aims to unify it under one solution that is flexible enough for all the different applications.

World#

The central datastructure for interaction with a scene is the semantic_world.world.World. The world is a mediator for bodies and their connections. It handles the validation of the world’s kinematic structure and the communication between the objects.

Physical Objects can be spawned by constructing a semantic_world.world.Body and a kinematic chain of those elements is added by specifying a semantic_world.world.Connection between bodies.

All those things have to be added to the world for full functionality. More information on the kinematic world model can be found here.

Views#

Views are aggregation objects for bodies and connections in the world. Semantically, a semantic_world.world.View is an interpretation of a set of links and connections. For instance, a Drawer can be seen as a view on a handle and a container that is connected via a fixed connection and where the container has some prismatic connection.

Views can be inferred by specifying rules that make up a view. TODO Bass stuff

Database#

The entire world can be saved to any database that has an sqlalchemy connector. The definitions and relationships for the database are automatically derived from the datastructures derived in the python package via the ormatic package. Since the datastructures for the forward calculations of the world are not defined compatibly, the types from semantic_world.spatial_types.spatial_types are mapped via JSON as columns. This is due to the fact, that this package uses casadi to speed up forward kinematics. The types for sqlalchemy are defined in semantic_world.orm.model. The interface to sqlalchemy is auto-generated to semantic_world.orm.ormatic_interface. The script to recreate the interface is found in here.

Contents