Learning Passive Policies

Riccardo Zanella; Federico Califano; Cristian Secchi; Stefano Stramigioli

doi:10.1007/978-3-031-76424-0_60

Learning Passive Policies

Riccardo Zanella^*, Federico Califano, Cristian Secchi, Stefano Stramigioli

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

2 Downloads (Pure)

Abstract

We merge the techniques of passivity-based control (PBC) and reinforcement learning (RL) in a robotic context, with the goal of learning passive control policies. We frame our contribution in a scenario where PBC is implemented by means of virtual energy tanks, a control technique developed to achieve closed-loop passivity for any arbitrary control input. The use of RL in combination with energy tanks allows to learn control policies which, under proper conditions, are structurally passive. Simulations show the validity of the approach, as well as novel research directions in energy-aware robotics.

Original language	English
Title of host publication	European Robotics Forum 2024 - 15th ERF
Editors	Cristian Secchi, Lorenzo Marconi
Publisher	Springer
Pages	338-343
Number of pages	6
Volume	1
ISBN (Print)	9783031764233
DOIs	https://doi.org/10.1007/978-3-031-76424-0_60
Publication status	Published - 1 Jan 2025
Event	European Robotics Forum, ERF 2024: ROBOTICS UNITES: People, Countries, Disciplines - Via della Fiera 23 – 47923 , Rimini, Italy Duration: 13 Mar 2024 → 15 Mar 2024 https://erf2024.eu/

Publication series

Name	Springer Proceedings in Advanced Robotics
Volume	32 SPAR
ISSN (Print)	2511-1256
ISSN (Electronic)	2511-1264

Conference

Conference	European Robotics Forum, ERF 2024
Abbreviated title	ERF 2024
Country/Territory	Italy
City	Rimini
Period	13/03/24 → 15/03/24
Internet address	https://erf2024.eu/

Keywords

2025 OA procedure
passivity-based control
reinforcement learning
energy tanks

Access to Document

10.1007/978-3-031-76424-0_60

978-3-031-76424-0_60Final published version, 285 KBLicence: Taverne

Cite this

@inproceedings{1f5ea8e2832e41c5847938a8bdef6b61,

title = "Learning Passive Policies",

abstract = "We merge the techniques of passivity-based control (PBC) and reinforcement learning (RL) in a robotic context, with the goal of learning passive control policies. We frame our contribution in a scenario where PBC is implemented by means of virtual energy tanks, a control technique developed to achieve closed-loop passivity for any arbitrary control input. The use of RL in combination with energy tanks allows to learn control policies which, under proper conditions, are structurally passive. Simulations show the validity of the approach, as well as novel research directions in energy-aware robotics.",

keywords = "2025 OA procedure, passivity-based control, reinforcement learning, energy tanks",

author = "Riccardo Zanella and Federico Califano and Cristian Secchi and Stefano Stramigioli",

note = "Publisher Copyright: {\textcopyright} The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.; European Robotics Forum, ERF 2024 : ROBOTICS UNITES: People, Countries, Disciplines, ERF 2024 ; Conference date: 13-03-2024 Through 15-03-2024",

year = "2025",

month = jan,

day = "1",

doi = "10.1007/978-3-031-76424-0_60",

language = "English",

isbn = "9783031764233",

volume = "1",

series = "Springer Proceedings in Advanced Robotics",

publisher = "Springer",

pages = "338--343",

editor = "Cristian Secchi and Lorenzo Marconi",

booktitle = "European Robotics Forum 2024 - 15th ERF",

address = "Germany",

url = "https://erf2024.eu/",

}

Learning Passive Policies. / Zanella, Riccardo; Califano, Federico; Secchi, Cristian et al.
European Robotics Forum 2024 - 15th ERF. ed. / Cristian Secchi; Lorenzo Marconi. Vol. 1 Springer, 2025. p. 338-343 (Springer Proceedings in Advanced Robotics; Vol. 32 SPAR).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Learning Passive Policies

AU - Zanella, Riccardo

AU - Califano, Federico

AU - Secchi, Cristian

AU - Stramigioli, Stefano

N1 - Publisher Copyright: © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

PY - 2025/1/1

Y1 - 2025/1/1

N2 - We merge the techniques of passivity-based control (PBC) and reinforcement learning (RL) in a robotic context, with the goal of learning passive control policies. We frame our contribution in a scenario where PBC is implemented by means of virtual energy tanks, a control technique developed to achieve closed-loop passivity for any arbitrary control input. The use of RL in combination with energy tanks allows to learn control policies which, under proper conditions, are structurally passive. Simulations show the validity of the approach, as well as novel research directions in energy-aware robotics.

AB - We merge the techniques of passivity-based control (PBC) and reinforcement learning (RL) in a robotic context, with the goal of learning passive control policies. We frame our contribution in a scenario where PBC is implemented by means of virtual energy tanks, a control technique developed to achieve closed-loop passivity for any arbitrary control input. The use of RL in combination with energy tanks allows to learn control policies which, under proper conditions, are structurally passive. Simulations show the validity of the approach, as well as novel research directions in energy-aware robotics.

KW - 2025 OA procedure

KW - passivity-based control

KW - reinforcement learning

KW - energy tanks

UR - http://www.scopus.com/inward/record.url?scp=85215811018&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-76424-0_60

DO - 10.1007/978-3-031-76424-0_60

M3 - Conference contribution

AN - SCOPUS:85215811018

SN - 9783031764233

VL - 1

T3 - Springer Proceedings in Advanced Robotics

SP - 338

EP - 343

BT - European Robotics Forum 2024 - 15th ERF

A2 - Secchi, Cristian

A2 - Marconi, Lorenzo

PB - Springer

T2 - European Robotics Forum, ERF 2024

Y2 - 13 March 2024 through 15 March 2024

ER -

Learning Passive Policies

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this