Ontable

Ontable is a term used in automated planning and robotics to denote that an object is resting directly on the surface of a table. It is commonly represented as a predicate, such as ontable(x), in planning formalisms like STRIPS and PDDL, where it helps describe the position of objects in a scene or problem instance. The predicate is independent of whether other objects are placed atop the object on the table; an object that is ontable may still have other objects on top of it.

Semantics and usage context

In many planning domains, ontable(x) expresses that the bottom of object x is in contact with the table. This means x is not supported by another object, though it may support other objects on top of it (for example, a block on the table with another block on top of it can still be ontable). Ontable is often used alongside predicates like on(x, y) to distinguish objects directly on the table from those on other objects, and alongside clear(x) to indicate that no object rests on top of x.

Examples

In Blocks World-like problems, a typical initial state might include ontable(A) and ontable(B) with clear(A) and clear(B), meaning A and B lie on the table with nothing on top of either. A possible goal could require on(A, B) while ontable(C) holds, specifying that A is placed on B and C remains on the table.

In robotics and perception, ontable appears in datasets and annotations to describe objects located on a tabletop surface, informing tasks such as grasping, sorting, and manipulation planning.

Relation to other concepts

Ontable is part of a family of location predicates that also includes on(x, y) for objects positioned on other objects, and clear(x) for objects without anything resting on them. Together, these predicates enable structured reasoning about tabletop scenes in both symbolic planning and robotic manipulation.