Hybrid Hessians for flexible optimization of pose graphs

Matthew Koichi Grimes, Dragomir Anguelov, Yann LeCun

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

Abstract

We present a novel "hybrid Hessian" six-degrees-of-freedom simultaneous localization and mapping (SLAM) algorithm. Our method allows for the smooth trade-off of accuracy for efficiency and for the incorporation of GPS measurements during real-time operation, thereby offering significant advantages over other SLAM solvers. Like other stochastic SLAM methods, such as SGD and TORO, our technique is robust to local minima and eliminates the need for costly relinearizations of the map. Unlike other stochastic methods, but similar to exact solvers, such as iSAM, our technique is able to process position-only constraints, such as GPS measurements, without introducing systematic distortions in the map. We present results from the Google Street View database, and compare our method with results from TORO. We show that our solver is able to achieve higher accuracy while operating within real-time bounds. In addition, as far as we are aware, this is the first stochastic SLAM solver capable of processing GPS constraints in real-time.

Original languageEnglish (US)
Title of host publicationIEEE/RSJ 2010 International Conference on Intelligent Robots and Systems, IROS 2010 - Conference Proceedings
Pages2997-3004
Number of pages8
DOIs
StatePublished - 2010
Event23rd IEEE/RSJ 2010 International Conference on Intelligent Robots and Systems, IROS 2010 - Taipei, Taiwan, Province of China
Duration: Oct 18 2010Oct 22 2010

Other

Other23rd IEEE/RSJ 2010 International Conference on Intelligent Robots and Systems, IROS 2010
CountryTaiwan, Province of China
CityTaipei
Period10/18/1010/22/10

Fingerprint

Global positioning system
Processing

ASJC Scopus subject areas

  • Artificial Intelligence
  • Human-Computer Interaction
  • Control and Systems Engineering

Cite this

Grimes, M. K., Anguelov, D., & LeCun, Y. (2010). Hybrid Hessians for flexible optimization of pose graphs. In IEEE/RSJ 2010 International Conference on Intelligent Robots and Systems, IROS 2010 - Conference Proceedings (pp. 2997-3004). [5650091] https://doi.org/10.1109/IROS.2010.5650091

Hybrid Hessians for flexible optimization of pose graphs. / Grimes, Matthew Koichi; Anguelov, Dragomir; LeCun, Yann.

IEEE/RSJ 2010 International Conference on Intelligent Robots and Systems, IROS 2010 - Conference Proceedings. 2010. p. 2997-3004 5650091.

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

Grimes, MK, Anguelov, D & LeCun, Y 2010, Hybrid Hessians for flexible optimization of pose graphs. in IEEE/RSJ 2010 International Conference on Intelligent Robots and Systems, IROS 2010 - Conference Proceedings., 5650091, pp. 2997-3004, 23rd IEEE/RSJ 2010 International Conference on Intelligent Robots and Systems, IROS 2010, Taipei, Taiwan, Province of China, 10/18/10. https://doi.org/10.1109/IROS.2010.5650091
Grimes MK, Anguelov D, LeCun Y. Hybrid Hessians for flexible optimization of pose graphs. In IEEE/RSJ 2010 International Conference on Intelligent Robots and Systems, IROS 2010 - Conference Proceedings. 2010. p. 2997-3004. 5650091 https://doi.org/10.1109/IROS.2010.5650091
Grimes, Matthew Koichi ; Anguelov, Dragomir ; LeCun, Yann. / Hybrid Hessians for flexible optimization of pose graphs. IEEE/RSJ 2010 International Conference on Intelligent Robots and Systems, IROS 2010 - Conference Proceedings. 2010. pp. 2997-3004
@inproceedings{42928628f46249809617e399dc0b8ac1,
title = "Hybrid Hessians for flexible optimization of pose graphs",
abstract = "We present a novel {"}hybrid Hessian{"} six-degrees-of-freedom simultaneous localization and mapping (SLAM) algorithm. Our method allows for the smooth trade-off of accuracy for efficiency and for the incorporation of GPS measurements during real-time operation, thereby offering significant advantages over other SLAM solvers. Like other stochastic SLAM methods, such as SGD and TORO, our technique is robust to local minima and eliminates the need for costly relinearizations of the map. Unlike other stochastic methods, but similar to exact solvers, such as iSAM, our technique is able to process position-only constraints, such as GPS measurements, without introducing systematic distortions in the map. We present results from the Google Street View database, and compare our method with results from TORO. We show that our solver is able to achieve higher accuracy while operating within real-time bounds. In addition, as far as we are aware, this is the first stochastic SLAM solver capable of processing GPS constraints in real-time.",
author = "Grimes, {Matthew Koichi} and Dragomir Anguelov and Yann LeCun",
year = "2010",
doi = "10.1109/IROS.2010.5650091",
language = "English (US)",
isbn = "9781424466757",
pages = "2997--3004",
booktitle = "IEEE/RSJ 2010 International Conference on Intelligent Robots and Systems, IROS 2010 - Conference Proceedings",

}

TY - GEN

T1 - Hybrid Hessians for flexible optimization of pose graphs

AU - Grimes, Matthew Koichi

AU - Anguelov, Dragomir

AU - LeCun, Yann

PY - 2010

Y1 - 2010

N2 - We present a novel "hybrid Hessian" six-degrees-of-freedom simultaneous localization and mapping (SLAM) algorithm. Our method allows for the smooth trade-off of accuracy for efficiency and for the incorporation of GPS measurements during real-time operation, thereby offering significant advantages over other SLAM solvers. Like other stochastic SLAM methods, such as SGD and TORO, our technique is robust to local minima and eliminates the need for costly relinearizations of the map. Unlike other stochastic methods, but similar to exact solvers, such as iSAM, our technique is able to process position-only constraints, such as GPS measurements, without introducing systematic distortions in the map. We present results from the Google Street View database, and compare our method with results from TORO. We show that our solver is able to achieve higher accuracy while operating within real-time bounds. In addition, as far as we are aware, this is the first stochastic SLAM solver capable of processing GPS constraints in real-time.

AB - We present a novel "hybrid Hessian" six-degrees-of-freedom simultaneous localization and mapping (SLAM) algorithm. Our method allows for the smooth trade-off of accuracy for efficiency and for the incorporation of GPS measurements during real-time operation, thereby offering significant advantages over other SLAM solvers. Like other stochastic SLAM methods, such as SGD and TORO, our technique is robust to local minima and eliminates the need for costly relinearizations of the map. Unlike other stochastic methods, but similar to exact solvers, such as iSAM, our technique is able to process position-only constraints, such as GPS measurements, without introducing systematic distortions in the map. We present results from the Google Street View database, and compare our method with results from TORO. We show that our solver is able to achieve higher accuracy while operating within real-time bounds. In addition, as far as we are aware, this is the first stochastic SLAM solver capable of processing GPS constraints in real-time.

UR - http://www.scopus.com/inward/record.url?scp=78651517575&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=78651517575&partnerID=8YFLogxK

U2 - 10.1109/IROS.2010.5650091

DO - 10.1109/IROS.2010.5650091

M3 - Conference contribution

SN - 9781424466757

SP - 2997

EP - 3004

BT - IEEE/RSJ 2010 International Conference on Intelligent Robots and Systems, IROS 2010 - Conference Proceedings

ER -