2015

Model Predictive Path Integral Control using Covariance Variable Importance Sampling

Williams, Grady, Aldrich, Andrew, Theodorou, Evangelos

Understand

In this paper we develop a Model Predictive Path Integral (MPPI) control algorithm based on a generalized importance sampling scheme and perform parallel optimization via sampling using a Graphics Processing Unit (GPU).

  • The proposed generalized importance sampling scheme allows for changes in the drift and diffusion terms of stochastic diffusion processes and plays a significant role in the performance of the model predictive control algorithm.
  • We compare the proposed algorithm in simulation with a model predictive control version of differential dynamic programming.

Built on

  • Differential dynamic programming

    D. H. Jacobson and D. Q. Mayne · 1970

    Earlier work this paper cites.

  • Stochastic Differential Equations And Applications

    A. Friedman · 1975

    Earlier work this paper cites.

  • Brownian Motion and Stochastic Calculus (Graduate Texts in Mathematics)

    I. Karatzas and S. E. Shreve · 1991

    Earlier work this paper cites.

  • Optimal control and estimation

    R. F. Stengel · 1994

    Earlier work this paper cites.

  • Linear theory for control of nonlinear stochastic systems

    H. J. Kappen · 2005

    Earlier work this paper cites.

  • A generalized iterative lqg method for locally-optimal feedback control of constrained nonlinear stochastic systems

    E. Todorov and W. Li · 2005

    Earlier work this paper cites.

  • Controlled Markov processes and viscosity solutions

    W. H. Fleming and H. M. Soner · 2006

    Earlier work this paper cites.

Similar

  • The grasp multiple micro-uav testbed

    Nathan Michael, Daniel Mellinger, Quentin Lindsey, and Vijay Kumar · 2010

    Cited alongside, same era.

  • Reinforcement learning of full-body humanoid motor skills

    F. Stulp, J. Buchli, E. Theodorou, and S. Schaal · 2010

    Cited alongside, same era.

  • Stochastic differential dynamic programming

    E. Theodorou, Y. Tassa, and E. Todorov · 2010

    Cited alongside, same era.

  • A generalized path integral approach to reinforcement learning

    E. A. Theodorou, J. Buchli, and S. Schaal · 2010

    Cited alongside, same era.

  • Relative entropy and free energy dualities: Connections to path integral and kl control

    E.A. Theodorou and E. Todorov · 2012

    Cited alongside, same era.

  • Dynamics and Control of Drifting in Automobiles

    R.Y Hindiyeh · 2013

    Cited alongside, same era.

Then

  • Reinforcement learning and synergistic control of the act hand

    E. Rombokas, M. Malhotra, E.A. Theodorou, E. Todorov, and Y. Matsuoka · 2013

    Later among the works it cites.

  • Policy search for path integral control

    Vicenç Gómez, Hilbert J Kappen, Jan Peters, and Gerhard Neumann · 2014

    Later among the works it cites.

  • Gpu based path integral control with learned dynamics

    G. Williams, E. Rombokas, and T. Daniel · 2014

    Later among the works it cites.

  • Real-time stochastic optimal control for multi-agent quadrotor swarms

    Original

    Vicenç Gómez, Sep Thijssen, Hilbert J Kappen, Stephen Hailes, and Andrew Symington · 2015

    Closest in time.

  • Nonlinear stochastic control and information theoretic dualities: Connections, interdependencies and thermodynamic interpretations

    Evangelos A. Theodorou · 2015

    Closest in time.

  • Path integral control and state-dependent feedback

    Sep Thijssen and HJ Kappen · 2015

    Closest in time.

Beyond the bibliography

alphaXiv searches the wider corpus for related work and actual follow-ups.

Open on alphaXiv

alphaXiv is searching for related work…