Automated Planning Encodings for the Manipulation of Articulated Objects in 3D with Gravity

Riccardo Bertolucci, Alessio Capitanelli, Marco Maratea, Fulvio Mastrogiovanni, Mauro Vallati

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

The manipulation of articulated objects plays an important role in real-world robot tasks, both in home and industrial environments. A lot of attention has been devoted to the development of ad hoc approaches and algorithms for generating the sequence of movements the robot has to perform in order to manipulate the object. Such approaches can hardly generalise on different settings, and are usually focused on 2D manipulations.

In this paper we introduce a set of PDDL+ formulations for performing automated manipulation of articulated objects in a three-dimensional workspace by a dual-arm robot. Presented formulations differ in terms of how gravity is modelled, considering different trade-offs between modelling accuracy and planning performance, and between human-readability and parsability by planners. Our experimental analysis compares the formulations on a range of domain-independent planners, that aim at generating plans for allowing a dual-arm robot to manipulate articulated objects of different sizes. Validation is performed in simulation on a Baxter robot.
Original languageEnglish
Title of host publicationAI*IA 2019 – Advances in Artificial Intelligence
Subtitle of host publicationXVIIIth International Conference of the Italian Association for Artificial Intelligence, Rende, Italy, November 19–22, 2019, Proceedings
EditorsMario Alviano, Gianluigi Greco, Francesco Scarcello
Place of PublicationCham
PublisherSpringer International Publishing AG
Pages135-150
Number of pages16
VolumeLNAI 11946
Edition1st
ISBN (Electronic)9783030351663
ISBN (Print)9783030351656
DOIs
Publication statusPublished - 17 Nov 2019
Event18th International Conference of the Italian Association for Artificial Intelligence - University of Calabria, Rende, Italy
Duration: 19 Nov 201922 Nov 2019
Conference number: 18
https://aiia2019.mat.unical.it/

Publication series

NameLecture Notes in Computer Science
PublisherSpringer
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference18th International Conference of the Italian Association for Artificial Intelligence
Abbreviated titleAIIA 2019
CountryItaly
CityRende
Period19/11/1922/11/19
Internet address

Fingerprint

Gravitation
Robots
Planning

Cite this

Bertolucci, R., Capitanelli, A., Maratea, M., Mastrogiovanni, F., & Vallati, M. (2019). Automated Planning Encodings for the Manipulation of Articulated Objects in 3D with Gravity. In M. Alviano, G. Greco, & F. Scarcello (Eds.), AI*IA 2019 – Advances in Artificial Intelligence: XVIIIth International Conference of the Italian Association for Artificial Intelligence, Rende, Italy, November 19–22, 2019, Proceedings (1st ed., Vol. LNAI 11946, pp. 135-150). (Lecture Notes in Computer Science). Cham: Springer International Publishing AG. https://doi.org/10.1007/978-3-030-35166-3_10
Bertolucci, Riccardo ; Capitanelli, Alessio ; Maratea, Marco ; Mastrogiovanni, Fulvio ; Vallati, Mauro. / Automated Planning Encodings for the Manipulation of Articulated Objects in 3D with Gravity. AI*IA 2019 – Advances in Artificial Intelligence: XVIIIth International Conference of the Italian Association for Artificial Intelligence, Rende, Italy, November 19–22, 2019, Proceedings. editor / Mario Alviano ; Gianluigi Greco ; Francesco Scarcello. Vol. LNAI 11946 1st. ed. Cham : Springer International Publishing AG, 2019. pp. 135-150 (Lecture Notes in Computer Science).
@inproceedings{858deea37b3b4913a1c731dafd6cabbc,
title = "Automated Planning Encodings for the Manipulation of Articulated Objects in 3D with Gravity",
abstract = "The manipulation of articulated objects plays an important role in real-world robot tasks, both in home and industrial environments. A lot of attention has been devoted to the development of ad hoc approaches and algorithms for generating the sequence of movements the robot has to perform in order to manipulate the object. Such approaches can hardly generalise on different settings, and are usually focused on 2D manipulations.In this paper we introduce a set of PDDL+ formulations for performing automated manipulation of articulated objects in a three-dimensional workspace by a dual-arm robot. Presented formulations differ in terms of how gravity is modelled, considering different trade-offs between modelling accuracy and planning performance, and between human-readability and parsability by planners. Our experimental analysis compares the formulations on a range of domain-independent planners, that aim at generating plans for allowing a dual-arm robot to manipulate articulated objects of different sizes. Validation is performed in simulation on a Baxter robot.",
keywords = "Mixed discrete-continuous planning, Robotics application",
author = "Riccardo Bertolucci and Alessio Capitanelli and Marco Maratea and Fulvio Mastrogiovanni and Mauro Vallati",
year = "2019",
month = "11",
day = "17",
doi = "10.1007/978-3-030-35166-3_10",
language = "English",
isbn = "9783030351656",
volume = "LNAI 11946",
series = "Lecture Notes in Computer Science",
publisher = "Springer International Publishing AG",
pages = "135--150",
editor = "Mario Alviano and Gianluigi Greco and Francesco Scarcello",
booktitle = "AI*IA 2019 – Advances in Artificial Intelligence",
address = "Switzerland",
edition = "1st",

}

Bertolucci, R, Capitanelli, A, Maratea, M, Mastrogiovanni, F & Vallati, M 2019, Automated Planning Encodings for the Manipulation of Articulated Objects in 3D with Gravity. in M Alviano, G Greco & F Scarcello (eds), AI*IA 2019 – Advances in Artificial Intelligence: XVIIIth International Conference of the Italian Association for Artificial Intelligence, Rende, Italy, November 19–22, 2019, Proceedings. 1st edn, vol. LNAI 11946, Lecture Notes in Computer Science, Springer International Publishing AG, Cham, pp. 135-150, 18th International Conference of the Italian Association for Artificial Intelligence, Rende, Italy, 19/11/19. https://doi.org/10.1007/978-3-030-35166-3_10

Automated Planning Encodings for the Manipulation of Articulated Objects in 3D with Gravity. / Bertolucci, Riccardo; Capitanelli, Alessio; Maratea, Marco; Mastrogiovanni, Fulvio; Vallati, Mauro.

AI*IA 2019 – Advances in Artificial Intelligence: XVIIIth International Conference of the Italian Association for Artificial Intelligence, Rende, Italy, November 19–22, 2019, Proceedings. ed. / Mario Alviano; Gianluigi Greco; Francesco Scarcello. Vol. LNAI 11946 1st. ed. Cham : Springer International Publishing AG, 2019. p. 135-150 (Lecture Notes in Computer Science).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Automated Planning Encodings for the Manipulation of Articulated Objects in 3D with Gravity

AU - Bertolucci, Riccardo

AU - Capitanelli, Alessio

AU - Maratea, Marco

AU - Mastrogiovanni, Fulvio

AU - Vallati, Mauro

PY - 2019/11/17

Y1 - 2019/11/17

N2 - The manipulation of articulated objects plays an important role in real-world robot tasks, both in home and industrial environments. A lot of attention has been devoted to the development of ad hoc approaches and algorithms for generating the sequence of movements the robot has to perform in order to manipulate the object. Such approaches can hardly generalise on different settings, and are usually focused on 2D manipulations.In this paper we introduce a set of PDDL+ formulations for performing automated manipulation of articulated objects in a three-dimensional workspace by a dual-arm robot. Presented formulations differ in terms of how gravity is modelled, considering different trade-offs between modelling accuracy and planning performance, and between human-readability and parsability by planners. Our experimental analysis compares the formulations on a range of domain-independent planners, that aim at generating plans for allowing a dual-arm robot to manipulate articulated objects of different sizes. Validation is performed in simulation on a Baxter robot.

AB - The manipulation of articulated objects plays an important role in real-world robot tasks, both in home and industrial environments. A lot of attention has been devoted to the development of ad hoc approaches and algorithms for generating the sequence of movements the robot has to perform in order to manipulate the object. Such approaches can hardly generalise on different settings, and are usually focused on 2D manipulations.In this paper we introduce a set of PDDL+ formulations for performing automated manipulation of articulated objects in a three-dimensional workspace by a dual-arm robot. Presented formulations differ in terms of how gravity is modelled, considering different trade-offs between modelling accuracy and planning performance, and between human-readability and parsability by planners. Our experimental analysis compares the formulations on a range of domain-independent planners, that aim at generating plans for allowing a dual-arm robot to manipulate articulated objects of different sizes. Validation is performed in simulation on a Baxter robot.

KW - Mixed discrete-continuous planning

KW - Robotics application

U2 - 10.1007/978-3-030-35166-3_10

DO - 10.1007/978-3-030-35166-3_10

M3 - Conference contribution

SN - 9783030351656

VL - LNAI 11946

T3 - Lecture Notes in Computer Science

SP - 135

EP - 150

BT - AI*IA 2019 – Advances in Artificial Intelligence

A2 - Alviano, Mario

A2 - Greco, Gianluigi

A2 - Scarcello, Francesco

PB - Springer International Publishing AG

CY - Cham

ER -

Bertolucci R, Capitanelli A, Maratea M, Mastrogiovanni F, Vallati M. Automated Planning Encodings for the Manipulation of Articulated Objects in 3D with Gravity. In Alviano M, Greco G, Scarcello F, editors, AI*IA 2019 – Advances in Artificial Intelligence: XVIIIth International Conference of the Italian Association for Artificial Intelligence, Rende, Italy, November 19–22, 2019, Proceedings. 1st ed. Vol. LNAI 11946. Cham: Springer International Publishing AG. 2019. p. 135-150. (Lecture Notes in Computer Science). https://doi.org/10.1007/978-3-030-35166-3_10