Evolution of neural networks for helicopter control: Why modularity matters

Renzo De Nardi, Julian Togelius, Owen E. Holland, Simon M. Lucas

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

    Abstract

    The problem of the automatic development of controllers for vehicles for which the exact characteristics are not known is considered in the context of miniature helicopter flocking. A methodology is proposed in which neural network based controllers are evolved in a simulation using a dynamic model qualitatively similar to the physical helicopter. Several network architectures and evolutionary sequences are investigated, and two approaches are found that can evolve very competitive controllers. The division of the neural network into modules and of the task into incremental steps seems to be a precondition for success, and we analyse why this might be so.

    Original languageEnglish (US)
    Title of host publication2006 IEEE Congress on Evolutionary Computation, CEC 2006
    Pages1799-1806
    Number of pages8
    StatePublished - 2006
    Event2006 IEEE Congress on Evolutionary Computation, CEC 2006 - Vancouver, BC, Canada
    Duration: Jul 16 2006Jul 21 2006

    Other

    Other2006 IEEE Congress on Evolutionary Computation, CEC 2006
    CountryCanada
    CityVancouver, BC
    Period7/16/067/21/06

    Fingerprint

    Helicopter
    Modularity
    Helicopters
    Neural Networks
    Neural networks
    Controller
    Controllers
    Flocking
    Precondition
    Network Architecture
    Network architecture
    Dynamic models
    Division
    Dynamic Model
    Module
    Methodology
    Simulation

    ASJC Scopus subject areas

    • Artificial Intelligence
    • Software
    • Theoretical Computer Science

    Cite this

    De Nardi, R., Togelius, J., Holland, O. E., & Lucas, S. M. (2006). Evolution of neural networks for helicopter control: Why modularity matters. In 2006 IEEE Congress on Evolutionary Computation, CEC 2006 (pp. 1799-1806). [1688525]

    Evolution of neural networks for helicopter control : Why modularity matters. / De Nardi, Renzo; Togelius, Julian; Holland, Owen E.; Lucas, Simon M.

    2006 IEEE Congress on Evolutionary Computation, CEC 2006. 2006. p. 1799-1806 1688525.

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

    De Nardi, R, Togelius, J, Holland, OE & Lucas, SM 2006, Evolution of neural networks for helicopter control: Why modularity matters. in 2006 IEEE Congress on Evolutionary Computation, CEC 2006., 1688525, pp. 1799-1806, 2006 IEEE Congress on Evolutionary Computation, CEC 2006, Vancouver, BC, Canada, 7/16/06.
    De Nardi R, Togelius J, Holland OE, Lucas SM. Evolution of neural networks for helicopter control: Why modularity matters. In 2006 IEEE Congress on Evolutionary Computation, CEC 2006. 2006. p. 1799-1806. 1688525
    De Nardi, Renzo ; Togelius, Julian ; Holland, Owen E. ; Lucas, Simon M. / Evolution of neural networks for helicopter control : Why modularity matters. 2006 IEEE Congress on Evolutionary Computation, CEC 2006. 2006. pp. 1799-1806
    @inproceedings{3fb6d5a60797444ba26b3f684e22a572,
    title = "Evolution of neural networks for helicopter control: Why modularity matters",
    abstract = "The problem of the automatic development of controllers for vehicles for which the exact characteristics are not known is considered in the context of miniature helicopter flocking. A methodology is proposed in which neural network based controllers are evolved in a simulation using a dynamic model qualitatively similar to the physical helicopter. Several network architectures and evolutionary sequences are investigated, and two approaches are found that can evolve very competitive controllers. The division of the neural network into modules and of the task into incremental steps seems to be a precondition for success, and we analyse why this might be so.",
    author = "{De Nardi}, Renzo and Julian Togelius and Holland, {Owen E.} and Lucas, {Simon M.}",
    year = "2006",
    language = "English (US)",
    isbn = "0780394879",
    pages = "1799--1806",
    booktitle = "2006 IEEE Congress on Evolutionary Computation, CEC 2006",

    }

    TY - GEN

    T1 - Evolution of neural networks for helicopter control

    T2 - Why modularity matters

    AU - De Nardi, Renzo

    AU - Togelius, Julian

    AU - Holland, Owen E.

    AU - Lucas, Simon M.

    PY - 2006

    Y1 - 2006

    N2 - The problem of the automatic development of controllers for vehicles for which the exact characteristics are not known is considered in the context of miniature helicopter flocking. A methodology is proposed in which neural network based controllers are evolved in a simulation using a dynamic model qualitatively similar to the physical helicopter. Several network architectures and evolutionary sequences are investigated, and two approaches are found that can evolve very competitive controllers. The division of the neural network into modules and of the task into incremental steps seems to be a precondition for success, and we analyse why this might be so.

    AB - The problem of the automatic development of controllers for vehicles for which the exact characteristics are not known is considered in the context of miniature helicopter flocking. A methodology is proposed in which neural network based controllers are evolved in a simulation using a dynamic model qualitatively similar to the physical helicopter. Several network architectures and evolutionary sequences are investigated, and two approaches are found that can evolve very competitive controllers. The division of the neural network into modules and of the task into incremental steps seems to be a precondition for success, and we analyse why this might be so.

    UR - http://www.scopus.com/inward/record.url?scp=34547345840&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=34547345840&partnerID=8YFLogxK

    M3 - Conference contribution

    AN - SCOPUS:34547345840

    SN - 0780394879

    SN - 9780780394872

    SP - 1799

    EP - 1806

    BT - 2006 IEEE Congress on Evolutionary Computation, CEC 2006

    ER -