Adaptive computational SLAM incorporating strategies of exploration and path planning

Jacky Baltes, Da Wei Kung, Wei Yen Wang, Chen Chien Hsu

Research output: Contribution to journalArticle

Abstract

Simultaneous localization and mapping (SLAM) is a well-known and fundamental topic for autonomous robot navigation. Existing solutions include the FastSLAM family-based approaches which are based on Rao-Blackwellized particle filter. The FastSLAM methods slow down greatly when the number of landmarks becomes large. Furthermore, the FastSLAM methods use a fixed number of particles, which may result in either not enough algorithms to find a solution in complex domains or too many particles and hence wasted computation for simple domains. These issues result in reduced performance of the FastSLAM algorithms, especially on embedded devices with limited computational capabilities, such as commonly used on mobile robots. To ease the computational burden, this paper proposes a modified version of FastSLAM called Adaptive Computation SLAM (ACSLAM), where particles are predicted only by odometry readings, and are updated only when an expected measurement has a maximum likelihood. As for the states of landmarks, they are also updated by the maximum likelihood. Furthermore, ACSLAM uses the effective sample size (ESS) to adapt the number of particles for the next generation. Experimental results demonstrated that the proposed ACSLAM performed 40% faster than FastSLAM 2.0 and also has higher accuracy.

Original languageEnglish
Article numbere23
JournalKnowledge Engineering Review
DOIs
Publication statusAccepted/In press - 2019 Jan 1

Fingerprint

Motion planning
Maximum likelihood
Mobile robots
Navigation
Robots

ASJC Scopus subject areas

  • Software
  • Artificial Intelligence

Cite this

Adaptive computational SLAM incorporating strategies of exploration and path planning. / Baltes, Jacky; Kung, Da Wei; Wang, Wei Yen; Hsu, Chen Chien.

In: Knowledge Engineering Review, 01.01.2019.

Research output: Contribution to journalArticle

@article{4b56acf1fea0443d9b37b4ace2d06e3e,
title = "Adaptive computational SLAM incorporating strategies of exploration and path planning",
abstract = "Simultaneous localization and mapping (SLAM) is a well-known and fundamental topic for autonomous robot navigation. Existing solutions include the FastSLAM family-based approaches which are based on Rao-Blackwellized particle filter. The FastSLAM methods slow down greatly when the number of landmarks becomes large. Furthermore, the FastSLAM methods use a fixed number of particles, which may result in either not enough algorithms to find a solution in complex domains or too many particles and hence wasted computation for simple domains. These issues result in reduced performance of the FastSLAM algorithms, especially on embedded devices with limited computational capabilities, such as commonly used on mobile robots. To ease the computational burden, this paper proposes a modified version of FastSLAM called Adaptive Computation SLAM (ACSLAM), where particles are predicted only by odometry readings, and are updated only when an expected measurement has a maximum likelihood. As for the states of landmarks, they are also updated by the maximum likelihood. Furthermore, ACSLAM uses the effective sample size (ESS) to adapt the number of particles for the next generation. Experimental results demonstrated that the proposed ACSLAM performed 40{\%} faster than FastSLAM 2.0 and also has higher accuracy.",
author = "Jacky Baltes and Kung, {Da Wei} and Wang, {Wei Yen} and Hsu, {Chen Chien}",
year = "2019",
month = "1",
day = "1",
doi = "10.1017/S0269888919000183",
language = "English",
journal = "Knowledge Engineering Review",
issn = "0269-8889",
publisher = "Cambridge University Press",

}

TY - JOUR

T1 - Adaptive computational SLAM incorporating strategies of exploration and path planning

AU - Baltes, Jacky

AU - Kung, Da Wei

AU - Wang, Wei Yen

AU - Hsu, Chen Chien

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Simultaneous localization and mapping (SLAM) is a well-known and fundamental topic for autonomous robot navigation. Existing solutions include the FastSLAM family-based approaches which are based on Rao-Blackwellized particle filter. The FastSLAM methods slow down greatly when the number of landmarks becomes large. Furthermore, the FastSLAM methods use a fixed number of particles, which may result in either not enough algorithms to find a solution in complex domains or too many particles and hence wasted computation for simple domains. These issues result in reduced performance of the FastSLAM algorithms, especially on embedded devices with limited computational capabilities, such as commonly used on mobile robots. To ease the computational burden, this paper proposes a modified version of FastSLAM called Adaptive Computation SLAM (ACSLAM), where particles are predicted only by odometry readings, and are updated only when an expected measurement has a maximum likelihood. As for the states of landmarks, they are also updated by the maximum likelihood. Furthermore, ACSLAM uses the effective sample size (ESS) to adapt the number of particles for the next generation. Experimental results demonstrated that the proposed ACSLAM performed 40% faster than FastSLAM 2.0 and also has higher accuracy.

AB - Simultaneous localization and mapping (SLAM) is a well-known and fundamental topic for autonomous robot navigation. Existing solutions include the FastSLAM family-based approaches which are based on Rao-Blackwellized particle filter. The FastSLAM methods slow down greatly when the number of landmarks becomes large. Furthermore, the FastSLAM methods use a fixed number of particles, which may result in either not enough algorithms to find a solution in complex domains or too many particles and hence wasted computation for simple domains. These issues result in reduced performance of the FastSLAM algorithms, especially on embedded devices with limited computational capabilities, such as commonly used on mobile robots. To ease the computational burden, this paper proposes a modified version of FastSLAM called Adaptive Computation SLAM (ACSLAM), where particles are predicted only by odometry readings, and are updated only when an expected measurement has a maximum likelihood. As for the states of landmarks, they are also updated by the maximum likelihood. Furthermore, ACSLAM uses the effective sample size (ESS) to adapt the number of particles for the next generation. Experimental results demonstrated that the proposed ACSLAM performed 40% faster than FastSLAM 2.0 and also has higher accuracy.

UR - http://www.scopus.com/inward/record.url?scp=85076157726&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85076157726&partnerID=8YFLogxK

U2 - 10.1017/S0269888919000183

DO - 10.1017/S0269888919000183

M3 - Article

AN - SCOPUS:85076157726

JO - Knowledge Engineering Review

JF - Knowledge Engineering Review

SN - 0269-8889

M1 - e23

ER -