ALAN: adaptive learning for multi-agent navigation

Julio Godoy, Tiannan Chen, Stephen J. Guy, Ioannis Karamouzas, Maria Gini

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

In multi-agent navigation, agents need to move towards their goal locations while avoiding collisions with other agents and obstacles, often without communication. Existing methods compute motions that are locally optimal but do not account for the aggregated motions of all agents, producing inefficient global behavior especially when agents move in a crowded space. In this work, we develop a method that allows agents to dynamically adapt their behavior to their local conditions. We formulate the multi-agent navigation problem as an action-selection problem and propose an approach, ALAN, that allows agents to compute time-efficient and collision-free motions. ALAN is highly scalable because each agent makes its own decisions on how to move, using a set of velocities optimized for a variety of navigation tasks. Experimental results show that agents using ALAN, in general, reach their destinations faster than using ORCA, a state-of-the-art collision avoidance framework, and two other navigation models.

Original languageEnglish (US)
Pages (from-to)1543-1562
Number of pages20
JournalAutonomous Robots
Volume42
Issue number8
DOIs
StatePublished - Dec 1 2018

Fingerprint

Navigation
Collision avoidance
Communication

Keywords

  • Action selection
  • Multi-agent coordination
  • Multi-agent navigation
  • Online learning

Cite this

ALAN : adaptive learning for multi-agent navigation. / Godoy, Julio; Chen, Tiannan; Guy, Stephen J.; Karamouzas, Ioannis; Gini, Maria.

In: Autonomous Robots, Vol. 42, No. 8, 01.12.2018, p. 1543-1562.

Research output: Contribution to journalArticle

Godoy, Julio ; Chen, Tiannan ; Guy, Stephen J. ; Karamouzas, Ioannis ; Gini, Maria. / ALAN : adaptive learning for multi-agent navigation. In: Autonomous Robots. 2018 ; Vol. 42, No. 8. pp. 1543-1562.
@article{16b1b9946c364e91a08050557a747b2b,
title = "ALAN: adaptive learning for multi-agent navigation",
abstract = "In multi-agent navigation, agents need to move towards their goal locations while avoiding collisions with other agents and obstacles, often without communication. Existing methods compute motions that are locally optimal but do not account for the aggregated motions of all agents, producing inefficient global behavior especially when agents move in a crowded space. In this work, we develop a method that allows agents to dynamically adapt their behavior to their local conditions. We formulate the multi-agent navigation problem as an action-selection problem and propose an approach, ALAN, that allows agents to compute time-efficient and collision-free motions. ALAN is highly scalable because each agent makes its own decisions on how to move, using a set of velocities optimized for a variety of navigation tasks. Experimental results show that agents using ALAN, in general, reach their destinations faster than using ORCA, a state-of-the-art collision avoidance framework, and two other navigation models.",
keywords = "Action selection, Multi-agent coordination, Multi-agent navigation, Online learning",
author = "Julio Godoy and Tiannan Chen and Guy, {Stephen J.} and Ioannis Karamouzas and Maria Gini",
year = "2018",
month = "12",
day = "1",
doi = "10.1007/s10514-018-9719-4",
language = "English (US)",
volume = "42",
pages = "1543--1562",
journal = "Autonomous Robots",
issn = "0929-5593",
publisher = "Springer Netherlands",
number = "8",

}

TY - JOUR

T1 - ALAN

T2 - adaptive learning for multi-agent navigation

AU - Godoy, Julio

AU - Chen, Tiannan

AU - Guy, Stephen J.

AU - Karamouzas, Ioannis

AU - Gini, Maria

PY - 2018/12/1

Y1 - 2018/12/1

N2 - In multi-agent navigation, agents need to move towards their goal locations while avoiding collisions with other agents and obstacles, often without communication. Existing methods compute motions that are locally optimal but do not account for the aggregated motions of all agents, producing inefficient global behavior especially when agents move in a crowded space. In this work, we develop a method that allows agents to dynamically adapt their behavior to their local conditions. We formulate the multi-agent navigation problem as an action-selection problem and propose an approach, ALAN, that allows agents to compute time-efficient and collision-free motions. ALAN is highly scalable because each agent makes its own decisions on how to move, using a set of velocities optimized for a variety of navigation tasks. Experimental results show that agents using ALAN, in general, reach their destinations faster than using ORCA, a state-of-the-art collision avoidance framework, and two other navigation models.

AB - In multi-agent navigation, agents need to move towards their goal locations while avoiding collisions with other agents and obstacles, often without communication. Existing methods compute motions that are locally optimal but do not account for the aggregated motions of all agents, producing inefficient global behavior especially when agents move in a crowded space. In this work, we develop a method that allows agents to dynamically adapt their behavior to their local conditions. We formulate the multi-agent navigation problem as an action-selection problem and propose an approach, ALAN, that allows agents to compute time-efficient and collision-free motions. ALAN is highly scalable because each agent makes its own decisions on how to move, using a set of velocities optimized for a variety of navigation tasks. Experimental results show that agents using ALAN, in general, reach their destinations faster than using ORCA, a state-of-the-art collision avoidance framework, and two other navigation models.

KW - Action selection

KW - Multi-agent coordination

KW - Multi-agent navigation

KW - Online learning

UR - http://www.scopus.com/inward/record.url?scp=85042177833&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85042177833&partnerID=8YFLogxK

U2 - 10.1007/s10514-018-9719-4

DO - 10.1007/s10514-018-9719-4

M3 - Article

AN - SCOPUS:85042177833

VL - 42

SP - 1543

EP - 1562

JO - Autonomous Robots

JF - Autonomous Robots

SN - 0929-5593

IS - 8

ER -