Dmitry Trifonov

    Miusskaya st. 4, Moscow, 125047, Russia
    Keldysh Institute of Applied Mathematics RAS


    Koshmanova N. P., Pavlovsky V. E., Trifonov  D. S.
    Reinforcement learning for manipulator control
    2012, Vol. 8, No. 4, pp.  689-704
    We present method for constructing manipulator control system with reinforcement learning algorithm. We construct learning algorithm which uses information about performed actions and their quality with respect to desired behaviour called «reward». The goal of the learning algorithm is to construct control system maximizing total reward. Learning algorithm and constructed control system were tested on the manipulator collision avoidance problem.
    Keywords: reinforcement learning, manipulator, control, newton-euler algorithm
    Citation: Koshmanova N. P., Pavlovsky V. E., Trifonov  D. S.,  Reinforcement learning for manipulator control, Rus. J. Nonlin. Dyn., 2012, Vol. 8, No. 4, pp.  689-704

    Back to the list