=Paper=
{{Paper
|id=Vol-2708/robontics8
|storemode=property
|title=Robot Meets World
|pdfUrl=https://ceur-ws.org/Vol-2708/robontics8.pdf
|volume=Vol-2708
|authors=Jona Thai,Michael Grüninger
|dblpUrl=https://dblp.org/rec/conf/jowo/ThaiG20
}}
==Robot Meets World==
https://ceur-ws.org/Vol-2708/robontics8.pdf
Robot Meets World1
Jona THAI a , Michael GRÜNINGER a
a Department of Mechanical and Industrial Engineering, University of Toronto, Ontario,
Canada M5S 3G8
Abstract. The representation of the world in which an autonomous robot interacts
with other objects within the world requires the integration of perception, reason-
ing, and action. The challenge of designing a suite of ontologies that are expres-
sive enough to support perception and reasoning is a preeminent challenge for the
knowledge representation and applied ontology communities. In this paper, we pro-
pose a series of challenge problems oriented around the ways in which a robot en-
counters the physical world. We outline an approach to the design and application
of ontologies that can be used to represent the problems and their solution.
Keywords. ontology process mereotopology robotics first-order logic PSL
1. Introduction
Imagine waking up in a well-lit, unrecognizable room. Although terrifying and stress-
inducing, a quick scan should allow you to quickly devise a plan of action. You are
able to discern what sharp objects to avoid, what objects are movable, ”free space” and
what would lead to a ”free fall”. Much of these common-sense reasoning is accumu-
lated through trial-and-error over the years. However, even more of this common-sense
is taught to us – for instance, we don’t need to get stabbed by a knife ourselves to un-
derstand that it is fatal. A small portion of our intelligence is genetic – for instance, we
figure out how to walk and grab things without explicit teaching or extensive experience.
This combination of spatial, temporal and physical knowledge allows us to reason about
the properties of spaces and objects that we have never encountered. The question is,
what is the minimum(albeit non-exhaustive) knowledge needed to represent a robot as
an embodied cognitive agent, one that possesses a physical form, is aware of it, and in-
teracts with other physical objects and other embodied cognitive agents? To this end, we
will explore several scenarios and propose relevant ontologies.
2. Informal Specification of Reasoning Problems for Robotics
Assuming a robot has successfully replicated our spatiotemporal and physical reasoning
capability, what is the minimum we can expect it to be able to do? In this section, we
present an initial set of reasoning problems that can be used to determine the suite of
ontologies that we will need.
1 Copyright © 2020 for this paper by its authors. Use permitted under Creative Commons License Attribution
4.0 International (CC BY 4.0).
Problem 1 Path Planning
Given a room containing a set of objects, find a path that the robot can take from one
location to another which avoids contact with any objects.
Problem 2 Relating Self to Object
Given an Article A, and a room containing a set of objects, find a path the robot can take
while carrying Article A from one location to another, while avoiding contact with any
objects.
Problem 3 Up, Down and All Around
Given a tall Obstacle A, wide Obstacle B and ledge C, find the optimal path the robot
can take.
Problem 4 Bob The Builder
Given a set of instructions and a set of objects, the robot should build a composite object.
Problem 5 Not Whole But Hole
Given a room with holes, find the optimal path the robot can take to the other side of the
room.
Problem 6 Tool Building
Given a room with a set of objects, and a chasm, assemble the objects in a way that the
robot can cross the chasm without falling into it.
Problem 7 Path Planning 2
Given a room with a set of moving objects, find a path the robot can take from one
location to another which avoids contact with any objects.
Problem 8 Breaking Barriers
Given a room with a set of objects forming a wall, move the objects or the robot in such
a way that it can traverse to the other side.
Problem 9 Lock(e) and Key
Given a Lock e, and a two vastly different keys, the robot must choose the right key the
first time.
Problem 10 Weigh the Way
Given a chasm and several objects of varying length and weight capacity, the robot needs
to accurately pick objects that will aid it in successfully crossing the chasm.
Problem 11 Together is Better
Given a wall W that is too tall for a single robot to traverse without assistance, find a
way for two robots to traverse the wall.
We believe that these problems address the minimum ontological commitments for a
robot to replicate our ability to reason about space and movement. The next step is to
design the weakest ontology possible for the robot to reason about solutions to the above
set of problems.
3. Existing/Related Work
Organizing knowledge around robotics is not a novel pursuit. Below we explore three
largely known ontologies within the field of robotics.
3.1. Core Ontology for Robotics and Automation(CORA)
CORA is an ontology specifically focused on the fields of industrial robotics, service
robotics, and autonomous robotics. It formally specifies mechanical components of
robots and robotics, types of robots and types of robot interactions. This is ideal for clas-
sifying different types of robotic designs and movements. This is achieved through an
alignment with upper ontologies such as SUMO, and OWL. [1]
3.2. KnowRob(2)
KnowRob is notable for its contributions in providing a heterogenous approach to seman-
tics in robotics. It utilizes a trove of ”narrative enabled episodic memories”, or NEEM,
for robots to learn from. This data is collected from multiple robotic systems, and stored
as RDF triples which are queried over the OpenEASE platform. Although effective, this
combination of software and ontologies is difficult to re-use and replicate, due to barriers
of entry of programming language and platform. [2]
3.3. OpenRobots Common Sense Ontology(ORO)
The OpenRobots Common Sense Ontology is closely related to KnowRob, both in shar-
ing of concepts and design concept. However, it differs from KnowRob in that it focuses
on human-robot interaction. It is aligned with the now defunct OpenCyc upper ontology.
It achieves its purpose by developing two main ontologiesn- one focusing on the theory
of the mind and one of ”common-sense”. The former helps develop context and make
sense of ambiguity in instructions, whereas the latter promotes an ability to reason about
actions to take. As such, ORO’s goal differs slightly from our proposed problem, as we
focus on problems in terms of robot-physical world interaction. [3]
3.4. What Ontologies Do We Need
Given the combined capability of these ontologies, one may wonder why we are propos-
ing yet another ontology for robotics. However, it should be noted that in terms of ap-
plication, we differ from ORO, in the sense that we focus on robot-physical world inter-
actions as compared to robot-human interaction. Moreover, we differ from CORA in the
sense that we do not seek to create a robotics ontology, but rather ontologies for robots
(that will support common-sense reasoning). In addition, all of the above ontologies do
not contain extensive first-order axiomatizations, which limits expressivity. KnowRob
cicumvents this through their hybrid, software-oriented approach. Although effective,
system complexity and different programming languages prove a barrier in reusability.
As such, our design of a reasoning framework purely based on first-order logic is novel.
How can we be sure that our set of designed ontologies are right for the application?
Following [4], an ontology works well if it does not generate unwanted models, and does
not constrict upon desired models. For example, mainstream designs of spatial reasoning
often utilize Euclidean 3D geometry. This intuition is not unfounded, as Newtonian me-
chanics occurs within this framework. However, this model is too strong, in the sense that
we do not need such formal geometry to represent overlapping regions. The same result
can be achieved with a mereotopology. At the same time, Euclidean geometry alone is
too weak for human-like spatial reasoning. This is because it does not contain axioms
that can differentiate between physical objects and the regions they occupy.
4. PSL Ontology
The PSL Ontology [5] is a modular first-order ontology that axiomatizes a set of intu-
itive semantic primitives that is adequate for describing the fundamental concepts of pro-
cesses. Two of the theories within the PSL Ontology play key roles in this paper – Tocctree
(theory of occurrence trees) and Tdisc state (theory of fluents and discrete change) 2 .
Within the PSL Ontology, an occurrence tree Γ is a partially ordered set of activity
occurrences, such that for a given set of activities, all discrete sequences of their occur-
rences are branches of the tree. An occurrence tree contains all occurrences of all atomic
activities; activity occurrences that are elements of an occurrence tree are called arboreal
activity occurrences.
Most applications of process ontologies are used to represent dynamic behaviour in
the world so that intelligent agents may make predictions about the future and explana-
tions about the past. In particular, these predictions and explanations are often concerned
with the state of the world and how that state changes. The PSL core theory Tdisc state is
intended to capture the basic intuitions about states and their relationship to activities.
Within the PSL Ontology, the notion of state is represented by reified fluents. Intu-
itively, a change in state is captured by fluents that are either achieved or falsified by an
activity occurrence. The prior relation is used to specify the fluents that are intuitively
true prior to an activity occurrence and the holds relation specifies the fluents that are
intuitively true after an activity occurrence.
Furthermore, a fluent can only be changed by the occurrence of activities. Thus, if
some fluent holds after an activity occurrence, but after an activity occurrence later along
the branch it is false, then an activity must occur at some point between that changes
the fluent. This also leads to the requirement that the fluents holding after an activity
occurrence will be the same fluents that are prior to any successor occurrence, since there
cannot be an activity occurring between them.
The key step in applying the PSL Ontology to the problems in Section 2 uses the
approach presented in [6], which begins with an ontology about a specific domain, and
then characterizes the ways in which activities can possibly change relationships among
objects in the domain. By identifying models of the domain ontology with states within
the process ontology, we can use properties of models to classify activities.
Axioms in the ontology are mapped to state constraints, which are universal sen-
tences containing only prior literals and a unique activity occurrence variable. For any
domain ontology T , there exists a set of state constraints known as the domain state on-
tology. The domain process ontology is designed by focussing on the relationship be-
2 This paper uses the axiomatization of the PSL Ontology found at
http://www.mel.nist.gov/psl/psl-ontology/.
tween models of the domain ontology and the states associated with activity occurrences
in the occurrence tree.
5. Formal Specification of reasoning problems
We can now provide a formal specification of the reasoning problems from Section 2,
and identify the ontologies that are required for the specification.
5.1. Revisiting Problem 1
The first step in the formalization of Problem 1 is to identify the basic ontological com-
mitments that are required to specify the problem.
We begin with the existence of a room that contains a set of objects. As observed
by [7], this leads to the distinction between physical objects and the spatial regions that
they occupy, and hence raises the approach of mereotopological pluralism – there is a
mereotopology on physical bodies and another distinct mereotopology on spatial regions.
Moreover, the notion of location requires a mereotopological harmony principle, that is,
the existence of a homomorphism from the mereotopology of physical objects to the
mereotopology of spatial regions. As a result of these requirements, we need to reuse the
Toccupy ontology, which the has signature
{region, part, physical body, physical part,C, physical C, occupies}
The next aspect of Problem 1 focusses on the statement ”that the robot can take from
one location to another”. In other words, we need an ontology that can represent how
objects can change location, that is, an object can occupy one region at one point in time,
and then occupy a different location at another point in time. The design of an ontology
that satisfies this requirement is based on the methodology of [6] and [8] by which a
domain ontology (in our current case being Toccupy ) is used to specify a domain state
ontology. In particular, axioms of the Occupy State Ontology include state constraints
such as
(∀x, y, r1 , r2 , o) prior(t physical part(x, y), o)∧ prior(loc(x, r1 ), o)∧ prior(loc(y, r2 ), o) ⊃ prior(t part(r1 , r2 ), o)
(1)
(∀x, y, o) prior(contact(x, y), o) ≡ (∃r1 , r2 ) prior(loc(x, r1 ), o)∧ prior(loc(y, r2 ), o)∧ prior(tC(r1 , r2 ), o)
(2)
The initial for the problem is specified as:
prior(loc(Robot, R1), Occinitial )prior(loc(Chair, R3), Occinitial ) (3)
while the goal state is:
(∃o) holds(loc(Robot, R4), o) (4)
The Motion Domain Process Ontology classifies activities by the ways in which objects can change
their location:
(∀a)motion(a) ≡ ((∀o)occurrence o f (o, a) ⊃ (∃x, r1 , r2 ) f alsi f ies(o, loc(x, r1 ))∧achieves(o, loc(x, r2 ))
(5)
The spatiotemporal objects that represent the path that the robot takes is axiomatized as a class of
complex activities within the PSL Ontology:
(∀a) path(a) ≡ ((∀a0 ) subactivity(a0 , a) ∧ primitive(a0 ) ⊃ motion(a0 )) (6)
Finally, the constraint in Problem 1 is axiomatized by:
(∀x, y, o, o1 , a) path(a)∧occurrence o f (o, a)∧subactivity occurrence(o1 , o) ⊃ ¬holds(contact(x, y), o)
(7)
5.2. Revisiting Problem 2
To identify which ontologies are needed to satisfy Problem 2, we can first attempt to
break down the question as per Problem 1. Again, we begin with a room containing a
set of objects, except with the addition of Article A. The next aspect of the problem then
explains the role of Article A - the robot has to get across the room, without bumping
any of the obstacles, while carrying Article A.
Clearly, this problem is an extension of Problem 1. In addition to understanding the
relation between an object and the physical space it occupies, the robot needs to now
understand the relation between itself and another object. Our approach to this problem
utilizes the mereotopology of connected objects proposed in [9], in which objects are
connected iff they can be summed together to form a new object. The idea is to consider
the robot and the object it is holding to constitute a new object that is the sum of the robot
and the object. In this way, a mereotopology is sufficient to represent the fundamental
ontological commitments for a robot to be holding an object.
5.3. 1, 2, 3...Infinity!
Perhaps evident from the previous two problems, the set of 12 competency questions
above build upon each other. Due to time and space constraints, we will not explore the
remaining questions in detail, but offer suggestions as to how these questions may be
viewed and tackled. For instance, Problem 5 and 9 require explicit formalization of non-
physical obstacles e.g. holes. To comprehend a hole, the robot needs to understand the
distinction between material and immaterial objects, perhaps through algebraic topol-
ogy. Problem 9 requires an additional ontology for shape to map complementary rela-
tions between concavity and convexity. Another example is Problem 10 and 11 - both of
these problems require the robot’s self-awareness regarding its own capabilities relative
to the obstacles before it. In addition, Problem 11 introduces the concept of collaboration
between robots, an ability of definite value in large scale problems e.g. construction.
6. Conclusion
A typical approach to the application of ontologies to the representation and solution of
commonsense reasoning problems is to use the strongest possible ontology (such as full
Euclidean geometry or ordered real closed fields) to represent the world in which a robot
is situated. In this paper we have taken a different perspective and sought the weakest
possible ontology that is sufficient to axiomatize the intended semantics presupposed by
the natural language statements of challenge problems for robotic knowledge represen-
tation and reasoning. In particular, we have seen how mereotopology, location, and pro-
cess ontologies are sufficient to axiomatize the fundamental ontological commitments of
robot path planning and limited interactions of robots with physical objects in their envi-
ronment. In the spirit of the Physical Turing Test [10], we have proposed a series of chal-
lenge problems that will serve as a platform for the design and evaluation of ontologies
to support the representation and solution of these problems.
References
[1] Prestes E, Carbonera JL, Rama Fiorini S, M Jorge VA, Abel M, Madhavan R, et al. Towards a core
ontology for robotics and automation. Robotics and Autonomous Systems. 2013;61(11):1193 – 1204.
Ubiquitous Robotics. Available from: http://www.sciencedirect.com/science/article/pii/
S0921889013000596.
[2] Beetz M, Beßler D, Haidu A, Pomarlan M, Bozcuoglu AK, Bartels G. Know Rob 2.0 — A 2nd Gen-
eration Knowledge Processing Framework for Cognition-Enabled Robotic Agents. 2018 IEEE Interna-
tional Conference on Robotics and Automation (ICRA). 2018:512–519.
[3] Lemaignan S, Ros R, Mösenlechner L, Alami R, Beetz M. ORO, a knowledge management platform
for cognitive architectures in robotics; 2010. p. 3548–3553.
[4] Aameri B, Gruninger M. Which Ontology is Right for Me? In: AAAI; 2015. .
[5] Gruninger M. Ontology of the Process Specification Language. In: Handbook on Ontologies in Infor-
mation Systems. Springer-Verlag; 2003. .
[6] Aameri B. Using Partial Automorphisms to Design Process Ontologies. In: Proceedings of the 7th
International Conference on Formal Ontology in Information Systems. IOS Press; 2012. p. 309–322.
[7] Aameri B, Grüninger M. Location Ontologies based on Mereotopological Pluralism. Applied Ontology.
2020;15(1):1–50.
[8] Aameri B, Gruninger M. A First-Order Axiomatization of Change in Mereotopology. In: Proceedings
of the 27th International Workshop on Qualitative Reasoning; 2013. p. 115–121.
[9] Gruninger M, Chui C, Ru Y, Thai J. A Mereology of Connected Structures. In: FOIS; 2020. .
[10] Ortiz CL. Why We Need a Physically Embodied Turing Test and What It Might Look Like. AI Maga-
zine. 2016;37(1):55–62.