Semantic maps - a multi-hierarchical model 64.425 Integrated - - PowerPoint PPT Presentation

▶
semantic maps a multi hierarchical model
SMART_READER_LITE
LIVE PREVIEW

Semantic maps - a multi-hierarchical model 64.425 Integrated - - PowerPoint PPT Presentation

Aug 10, 2023 180 likes •593 views

MIN-Fakult at Fachbereich Informatik Universit at Hamburg Semantic maps Semantic maps - a multi-hierarchical model 64.425 Integrated Seminar: Intelligent Robotics Paul Anton Universit at Hamburg Fakult at f ur Mathematik,

slide-1
SLIDE 1

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Semantic maps

Semantic maps - a multi-hierarchical model

64.425 Integrated Seminar: Intelligent Robotics Paul Anton

Universit¨ at Hamburg Fakult¨ at f¨ ur Mathematik, Informatik und Naturwissenschaften Fachbereich Informatik Technische Aspekte Multimodaler Systeme

  • 07. December 2015
  • P. Anton

1

slide-2
SLIDE 2

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Semantic maps

Outline

  • 1. Semantic knowledge
  • 2. Multi-hierarchical model
  • 3. Applications
  • 4. Critical Evaluation
  • 5. Bibliography
  • P. Anton

2

slide-3
SLIDE 3

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Semantic knowledge - Traditional robot maps Semantic maps

Traditional robot maps

Metric map showing lines extracted from laser range scans - [6]. Space segmented into topological nodes.

  • P. Anton

3

slide-4
SLIDE 4

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Semantic knowledge - Necessity Semantic maps

Semantic knowledge. . .

Empowers mobile robots

◮ reasoning capabilities ◮ autonomy ◮ enhanced mobility ◮ efficiency ◮ interaction ◮ communication skills ◮ . . .

  • P. Anton

4

slide-5
SLIDE 5

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Semantic knowledge - Necessity Semantic maps

Semantic knowledge. . .

Empowers mobile robots

◮ reasoning capabilities ◮ autonomy ◮ enhanced mobility ◮ efficiency ◮ interaction ◮ communication skills ◮ . . .

  • P. Anton

4

slide-6
SLIDE 6

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Semantic knowledge - Necessity Semantic maps

Semantic knowledge. . .

Empowers mobile robots

◮ reasoning capabilities ◮ autonomy ◮ enhanced mobility ◮ efficiency ◮ interaction ◮ communication skills ◮ . . .

  • P. Anton

4

slide-7
SLIDE 7

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Semantic knowledge - Necessity Semantic maps

Semantic knowledge. . .

Empowers mobile robots

◮ reasoning capabilities ◮ autonomy ◮ enhanced mobility ◮ efficiency ◮ interaction ◮ communication skills ◮ . . .

  • P. Anton

4

slide-8
SLIDE 8

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Semantic knowledge - Necessity Semantic maps

Semantic knowledge. . .

Empowers mobile robots

◮ reasoning capabilities ◮ autonomy ◮ enhanced mobility ◮ efficiency ◮ interaction ◮ communication skills ◮ . . .

  • P. Anton

4

slide-9
SLIDE 9

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Semantic knowledge - Necessity Semantic maps

Semantic knowledge. . .

Empowers mobile robots

◮ reasoning capabilities ◮ autonomy ◮ enhanced mobility ◮ efficiency ◮ interaction ◮ communication skills ◮ . . .

  • P. Anton

4

slide-10
SLIDE 10

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Semantic knowledge - Necessity Semantic maps

Semantic knowledge. . .

Empowers mobile robots

◮ reasoning capabilities ◮ autonomy ◮ enhanced mobility ◮ efficiency ◮ interaction ◮ communication skills ◮ . . .

  • P. Anton

4

slide-11
SLIDE 11

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Semantic knowledge - Necessity Semantic maps

Semantic knowledge. . .

Empowers mobile robots

◮ reasoning capabilities ◮ autonomy ◮ enhanced mobility ◮ efficiency ◮ interaction ◮ communication skills ◮ . . .

  • P. Anton

4

slide-12
SLIDE 12

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Semantic knowledge - Properties and challenges Semantic maps

Properties and challenges

◮ co-exist with other components ◮ dealing with uncertain perceptions ◮ real-world indoor environment:

◮ dynamic ◮ appearance changes ◮ perception of environment

◮ properties of the sensors employed

  • P. Anton

5

slide-13
SLIDE 13

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Semantic knowledge - Properties and challenges Semantic maps

Properties and challenges

◮ co-exist with other components ◮ dealing with uncertain perceptions ◮ real-world indoor environment:

◮ dynamic ◮ appearance changes ◮ perception of environment

◮ properties of the sensors employed

  • P. Anton

5

slide-14
SLIDE 14

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Semantic knowledge - Properties and challenges Semantic maps

Properties and challenges

◮ co-exist with other components ◮ dealing with uncertain perceptions ◮ real-world indoor environment:

◮ dynamic ◮ appearance changes ◮ perception of environment

◮ properties of the sensors employed

  • P. Anton

5

slide-15
SLIDE 15

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Semantic knowledge - Properties and challenges Semantic maps

Properties and challenges

◮ co-exist with other components ◮ dealing with uncertain perceptions ◮ real-world indoor environment:

◮ dynamic ◮ appearance changes ◮ perception of environment

◮ properties of the sensors employed

  • P. Anton

5

slide-16
SLIDE 16

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Semantic knowledge - Properties and challenges Semantic maps

Properties and challenges

◮ co-exist with other components ◮ dealing with uncertain perceptions ◮ real-world indoor environment:

◮ dynamic ◮ appearance changes ◮ perception of environment

◮ properties of the sensors employed

  • P. Anton

5

slide-17
SLIDE 17

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Multi-hierarchical model Semantic maps

Multi-hierarchical semantic maps for mobile robotics

Galindo et. al.

”In this paper, we propose an approach to allow a mobile robot to build a semantic map from sensor data, and to use this semantic information in the performance of navigation tasks.“ - [3]

◮ Robot Task Planning using Semantic Maps [4] ◮ Monitoring the execution of robot plans using semantic

knowledge [2]

  • P. Anton

6

slide-18
SLIDE 18

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Multi-hierarchical model - Spatial hierarchy Semantic maps

Spatial hierarchy

Stores spatial and metric information of the environment

Spatial hierarchy - [3].

spatial environment the topology of the space images of objects and local grid maps

  • P. Anton

7

slide-19
SLIDE 19

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Multi-hierarchical model - Spatial hierarchy Semantic maps

Spatial hierarchy

Stores spatial and metric information of the environment

Spatial hierarchy - [3].

spatial environment the topology of the space images of objects and local grid maps

  • P. Anton

7

slide-20
SLIDE 20

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Multi-hierarchical model - Spatial hierarchy Semantic maps

Spatial hierarchy

Stores spatial and metric information of the environment

Spatial hierarchy - [3].

spatial environment the topology of the space images of objects and local grid maps

  • P. Anton

7

slide-21
SLIDE 21

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Multi-hierarchical model - Spatial hierarchy Semantic maps

Spatial hierarchy

Stores spatial and metric information of the environment

Spatial hierarchy - [3].

spatial environment the topology of the space images of objects and local grid maps

  • P. Anton

7

slide-22
SLIDE 22

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Multi-hierarchical model - Conceptual hierarchy Semantic maps

Conceptual hierarchy

Provides modelling of semantic knowledge and human-like inference capabilities

Conceptual hierarchy - [3].

common ancestor general categories specific concepts individual instances

  • P. Anton

8

slide-23
SLIDE 23

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Multi-hierarchical model - Conceptual hierarchy Semantic maps

Conceptual hierarchy

Provides modelling of semantic knowledge and human-like inference capabilities

Conceptual hierarchy - [3].

common ancestor general categories specific concepts individual instances

  • P. Anton

8

slide-24
SLIDE 24

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Multi-hierarchical model - Conceptual hierarchy Semantic maps

Conceptual hierarchy

Provides modelling of semantic knowledge and human-like inference capabilities

Conceptual hierarchy - [3].

common ancestor general categories specific concepts individual instances

  • P. Anton

8

slide-25
SLIDE 25

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Multi-hierarchical model - Conceptual hierarchy Semantic maps

Conceptual hierarchy

Provides modelling of semantic knowledge and human-like inference capabilities

Conceptual hierarchy - [3].

common ancestor general categories specific concepts individual instances

  • P. Anton

8

slide-26
SLIDE 26

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Multi-hierarchical model - Conceptual hierarchy Semantic maps

Conceptual hierarchy

Provides modelling of semantic knowledge and human-like inference capabilities

Conceptual hierarchy - [3].

common ancestor general categories specific concepts individual instances

  • P. Anton

8

slide-27
SLIDE 27

Linking via anchoring

Anchoring as a technique of connecting both hierarchies - [3].

slide-28
SLIDE 28

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Multi-hierarchical model - Description logics Semantic maps

Description logics

Define the relevant concepts (terminology) ↓ Specify properties of objects and entities (description) ↓ Represent the knowledge of an application domain (the world)

  • P. Anton

10

slide-29
SLIDE 29

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Multi-hierarchical model - Description logics Semantic maps

Description logics (cont.)

◮ There are individuals ◮ Connected through roles ◮ Concepts are sets of individuals

◮ simple concept: C

Space

◮ conjunction of concepts: C1 ⊓ C2

Area ≡ Space ⊓ (> 0hasDoor)

◮ disjunction of concepts: C1 ⊔ C2 ◮ negation of concepts: ¬C

Door ≡ ¬Window

◮ existential restriction: ∈ R.C

∈ hasBook.Bookcase

◮ universal restriction: ∀R.C

∀hasBed.Bedroom

  • P. Anton

11

slide-30
SLIDE 30

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Multi-hierarchical model - Description logics Semantic maps

Description logics (cont.)

ABox

◮ individual belongs to a class:

C(i)

◮ roles link individuals:

R(i,j) TBox

◮ generic/specific

C1 subclassOf C2

◮ equivalence

C1 equivalentClass C2

◮ disjunction

disjointWith(C1, C2)

  • P. Anton

12

slide-31
SLIDE 31

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Multi-hierarchical model - Description logics Semantic maps

Description logics (cont.)

Description Logic perspective of the semantic map - [4].

  • P. Anton

13

slide-32
SLIDE 32

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Multi-hierarchical model - Reasoning mechanisms Semantic maps

Reasoning mechanisms

◮ subsumption

KB C1 ⊑ C2

◮ equivalence

KB C1 ≡ C2

◮ instance checking

KB C(i)

  • P. Anton

14

slide-33
SLIDE 33

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Applications - Case Study Semantic maps

Case Study

Inferring Robot Goals from Semantic Knowledge [5]

What happens if the existing knowledge turns out to be in conflict with the robot’s observations?

  • 1. update the semantic knowledge base
  • 2. question the validity of its perceptions
  • 3. modify the environment
  • P. Anton

15

slide-34
SLIDE 34

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Applications - Case Study Semantic maps

Define normative relations and concepts

◮ Set of disjoint concepts

ρ = {P1, P2, . . . Pn}, i.e., ∀a, a ⊑ Pi ⇒ ∄j, j = i, a ⊑ Pj

◮ Define normative relations

Nr : NC → ρ ∀b ⊑ NC ⇒ ∃Pj ∈ ρ, b → [FILLS : NrPj]

◮ Separate normative concepts

NC = △ ∪ , △ ∩ = ∅

  • P. Anton

16

slide-35
SLIDE 35

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Applications - Case Study Semantic maps

Define normative relations and concepts (cont.)

Description logic interpretation of a domain[5].

  • P. Anton

17

slide-36
SLIDE 36

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Applications - Case Study Semantic maps

Norm violation detection

if ∃k ⊑ C, k → [FILLS : Nr, y], y ⊑ Pj ∈ ρ, Pj = Pi then y ⊑ Pj ∧ y ⊑ Pi incoherent

  • P. Anton

18

slide-37
SLIDE 37

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Critical Evaluation Semantic maps

Critical Evaluation

Representation of spatial knowledge

Manually built ontology. Issues: how to handle uncertainty?

Inference mechanism

Based on anchoring [1] - probabilistic inference engine

  • P. Anton

19

slide-38
SLIDE 38

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Critical Evaluation Semantic maps

Critical Evaluation (cont.)

Sources of semantic information

◮ objects ◮ . . .

Issues: scarcity of objects, reliable object categorization Solutions - [1]

◮ general appearance of places ◮ geometry of places ◮ topological structure

  • P. Anton

20

slide-39
SLIDE 39

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Semantic maps

Questions?

Thank you for your attention :)

  • P. Anton

21

slide-40
SLIDE 40

Universit¨ at Hamburg

MIN-Fakult¨ at Fachbereich Informatik Bibliography Semantic maps

Bibliography

[1] Andrzej Pronobis and Patric Jensfelt. Geometry And Topology, 2011. [2] Abdelbaki Bouguerra, Lars Karlsson, and Alessandro Saffiotti. Monitoring the execution of robot plans using semantic knowledge. Robotics and Autonomous Systems, 56(11):942–954, 2008. [3]

  • C. Galindo, a. Saffiotti, S. Coradeschi, P. Buschka, J. a. Fern´

andez-Madrigal, and J. Gonz´ alez. Multi-hierarchical semantic maps for mobile robotics. 2005 IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS, (3):3492–3497, 2005. [4] Cipriano Galindo, Juan Antonio Fern´ andez-Madrigal, Javier Gonz´ alez, and Alessandro Saffiotti. Robot task planning using semantic maps. Robotics and Autonomous Systems, 56(11):955–966, 2008. [5] Cipriano Galindo and Alessandro Saffiotti. Inferring robot goals from violations of semantic knowledge. Robotics and Autonomous Systems, 61(10):1131–1143, 2013. [6] H Zender, O Mart´ ınez Mozos, P Jensfelt, Geert-jan M Kruijff, and W Burgard. Conceptual Spatial Representations for Indoor Mobile Robots. Robotics and . . . , 6:493–502, 2008.

  • P. Anton

22