Scribd Logo
Upload

Welcome to Scribd!

  • 219 views

    Drake

    Uploaded by

    Robin Chicaiza
    Drake is a toolbox for modeling, simulating, and analyzing nonlinear dynamical systems. It was developed by researchers at MIT for applications such as robotics. Drake uses MATLAB and Simulink to provide algorithms for motion planning, control design, and stability analysis. It supports various robot models through URDF files and simulation through Bullet. Drake aims to exploit the mathematical structure of nonlinear systems for efficient analysis and control synthesis. The toolbox, along with documentation and examples, is freely available online.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    Drake

    Uploaded by

    Robin Chicaiza
    219 views27 pages
    Drake is a toolbox for modeling, simulating, and analyzing nonlinear dynamical systems. It was developed by researchers at MIT for applications such as robotics. Drake uses MATLAB and Simulink to provide algorithms for motion planning, control design, and stability analysis. It supports various robot models through URDF files and simulation through Bullet. Drake aims to exploit the mathematical structure of nonlinear systems for efficient analysis and control synthesis. The toolbox, along with documentation and examples, is freely available online.

    Original Description:

    Drake mit

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Drake is a toolbox for modeling, simulating, and analyzing nonlinear dynamical systems. It was developed by researchers at MIT for applications such as robotics. Drake uses MATLAB and Simulink to provide algorithms for motion planning, control design, and stability analysis. It supports various robot models through URDF files and simulation through Bullet. Drake aims to exploit the mathematical structure of nonlinear systems for efficient analysis and control synthesis. The toolbox, along with documentation and examples, is freely available online.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    219 views27 pages

    Drake

    Uploaded by

    Robin Chicaiza
    Drake is a toolbox for modeling, simulating, and analyzing nonlinear dynamical systems. It was developed by researchers at MIT for applications such as robotics. Drake uses MATLAB and Simulink to provide algorithms for motion planning, control design, and stability analysis. It supports various robot models through URDF files and simulation through Bullet. Drake aims to exploit the mathematical structure of nonlinear systems for efficient analysis and control synthesis. The toolbox, along with documentation and examples, is freely available online.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    You are on page 1of 27

    Drake: A planning, control,

    and analysis toolbox for


    nonlinear dynamical systems
    Russ Tedrake and Scott Kuindersma
    Robot Locomotion Group
    MIT CSAIL
    All of this is done using MATLAB

    and were giving away the code.


    What is Drake?
    Library for simulating, analyzing, and designing
    controllers for nonlinear dynamical systems

    Class hierarchy built on top of Simulink

    Exploits mathematical structure of nonlinear


    systems for design and analysis

    Used for research and teaching at MIT


    You could use Drake for
    Analyze the stability of your systems

    Design nonlinear feedback controllers for complicated


    (nonlinear, underactuated) dynamical systems

    Perform fast IK, trajectory and feedback-motion planning

    Compute invariant "funnels" along trajectories (derived


    from your own motion planning software) for robust motion
    planning

    Explore your own algorithms using Drake's


    implementations of model robot systems
    Why MATLAB?
    Rapid prototyping

    Easy interface to solvers

    SNOPT, Gurobi, Mosek, SeDuMi, CPLEX,


    CVX(GEN),

    Access to Simulink ODE solvers

    Easy visualization, GUIs, I/O, debugging

    Interaction with Java and C++ (via mex)


    Drake Features
    Many algorithms available for planning, verification,
    control, state estimation

    Support for URDFs

    Integration with Bullet for simulation and collision


    detection

    OpenGL visualizer

    Lots of examples: Atlas, PR2, fixed-wing aircraft,


    Structured Nonlinearity
    Simulink is a powerful tool
    combining and simulating
    dynamical systems

    Drake adds class structure for


    exploiting mathematical
    structure of nonlinear systems

    Important for modern nonlinear


    control synthesis
    Polynomial Dynamical Systems

    State space description:


    x = f (x, u, t)
    !
    y = g(x, u, t)
    where f and g are polynomial

    Verifying stability with Lyapunov functions can by


    done via convex optimization

    Region of attraction analysis via bilinear optimization


    Algebraic Rigid Body Dynamics

    Rigid body equations of motion can often be


    expressed as polynomials, e.g.,

    Change coordinates and add a constraint:


    Quadrotor ROA
    Verification with Inertia Variation of 10 Percent
    8
    Nominal Verified Basin
    6 Uncertain Verified Basin

    8
    0.1 0.08 0.06 0.04 0.02 0 0.02 0.04 0.06 0.08 0.1

    Quadrotor Example
    Simple Walking Models
    Flying Through Forests
    Anirudha Majumdar & Russ Tedrake
    t =0.00000 t =0.00000

    10 10

    8 8

    6 6

    4 4

    2 2

    0 0

    1 2 3 4 5 6 7 1 2 3 4 5 6 7
    Atlas Planning & Control
    Whole-body collision-free
    motion planning

    Footstep planning

    Dynamic whole-body control


    for locomotion and
    manipulation

    LCM interface to other


    software components
    Motion Planning

    Nonlinear optimization (via SNOPT)

    Large library of constraint classes for specifying


    problems

    Integration with Bullet for collision detection

    Fast: whole-body planning at interactive rates


    Walking Control

    Quadratic programming-based controller

    Significant speed requirement: exploits mex

    Custom efficient QP solver in Drake

    Exploits fast computation of dynamics, kinematics,


    and contact constraints (and derivatives)
    Atlas balancing video
    (demonstrates controller and motion planner)
    Online Education

    Underactuated Robotics course at MIT

    Upcoming edX course in Fall 2014


    Where to get it
    http://drake.mit.edu

    Support for Linux, Mac, and Windows

    Active and growing user community

    You might also like

    pFad - Phonifier reborn

    Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

    Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


    Alternative Proxies:

    Alternative Proxy

    pFad Proxy

    pFad v3 Proxy

    pFad v4 Proxy