Robotics
TypeDB’s ability to model complexity is extremely useful in robotics, where worlds, datasets and interactions can be extremely heterogenous.
The TypeDB Robotics open-source example shows how to use TypeDB for robotic navigation of the world. It is based on a real-world usage developed by Joris Sijs at the TNO University.
Installation
The robotics example is released as a schema and data file (note: good for small or medium-sized examples, but not recommended for large datasets!). You can easily load the latest schema and dataset, either using TypeDB Cloud to create a server with the dataset preloaded, or using TypeDB Console:
typdb console --username=<USERNAME> --address=<ADDRESS> --command="database create-init robotics http://github.com/typedb-osi/typedb-robotics/releases/latest/download/schema_floorplan.tql http://github.com/typedb-osi/typedb-robotics/releases/latest/download/data_floorplan_assignment.tql"Schema
This schema models a physical and mathematical domain, defining concepts ranging from physical objects and their spatial relationships to abstract mathematical constructs. The schema serves as a knowledge graph for describing, locating, and analyzing the components of a given environment, such as a building or a room. It is designed to be highly extensible and hierarchical, using abstract types to create a flexible and powerful model.
Core Concepts
Key Entities
The schema is built around two primary, top-level abstractions:
-
physical_entity: Represents anything that exists in the real world. This is the root for all concrete, tangible things. It owns attributes for existence, name, and id.
-
being: A living or artificial entity, like a human or a robot.
-
space: An open or closed area, such as a room or an open_space.
-
artifact: A manufactured item. This is further divided into stationary_artifact (like a building or wall) and constituent_artifact (like furniture or toys).
-
-
abstract_entity: Represents non-physical concepts, typically for defining and describing physical entities.
-
mathematical: Defines abstract shapes (shape), lines (line), and points (point), which can be used to model the physical world. For example, a room can have the shape of a box.
-
measure: Represents quantifiable features, like distance or height. These concepts are used to describe the properties of physical entities.
-
time: A simple entity to capture datetime values.
-
Key Relationships
The schema’s true power lies in its relations, which define how these concepts connect to each other. They are also categorized into mathematics and physics to reflect their purpose.
-
featuring: A fundamental relation that links a subject (a physical entity) to a feature (an abstract measure). This allows you to say, for example, a human has a height.
-
locating: A physical_entity is located in a space. This is a core relation for modeling the position of objects and beings within an environment.
-
taking_form: A mathematical concept is used to describe the form of a physical entity. For instance, a room is enclosed by a box, and a wall has the form of a rectangle.
-
positioning: Defines the precise spatial relationship between two points using attributes like lin_x and rot_x to represent coordinates and rotation.
-
composing: Defines a hierarchical structure, such as a building being composed of multiple rooms (volume) and walls (partition).
Key Attributes
The schema includes a rich set of attributes to describe entities and relations.
-
Existence: The exists and not_exists attributes, which are sub-types of the abstract existance, are used to model the current state of a physical_entity. This is useful for tracking things that may be created or destroyed.
-
Quantifiables: The schema has a comprehensive hierarchy of quantifiable attributes, ranging from continuous values like distance to categorical ones like color_name. This allows for highly detailed and specific data representation.