CS545_Lecture_13

CS545_Lecture_13 - CS545-Contents XIII Trajectory Planning...

Info icon This preview shows pages 1–7. Sign up to view the full content.

View Full Document Right Arrow Icon
CS545—Contents XIII Trajectory Planning Control Policies Desired Trajectories Optimization Methods Dynamical Systems Reading Assignment for Next Class See http://www-clmc.usc.edu/~cs545
Image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Learning Policies is the Goal of Learning Control Policy: u t ( ) = p x t ( ) , t , α ( )
Image of page 2
Dynamic Programming & Reinforcement Learning Movement System Nonlinear Controller (Policy) Desired Behavior u x Dynamic Programming requires a model of the movement system Reinforcement Learning can work without models of the movement system Essentials both techniques require to learn a high-dimensional “value function” that assesses the quality of an action u in a state x learning the value function is a complex nonstationary, nonlinear learning process both methods die the curse of dimensionality V = max u r x , u ( ) + τ V x ( ) x f x , u ( ) (HJB-Eqn.)
Image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Desired Trajectories Essentials prescribe a desired trajectory convert desired trajectory into a (time-dependent) control policy, e.g., by PD- controller Problems Where do desired trajectories come from How to accomplish reactive control How to generalize to new tasks or new situations θ , θ ( ) desired = f ξ initial , ξ target , t ( ) u = p x , t , α ( ) = k θ θ t ( ) desired θ ( ) + k θ θ t ( ) desired θ ( )
Image of page 4
Desired Trajectories (cont’d) There is a difference between PATH and TRAJECTORY planning A trajectory involves geometry AND time A path involves only geometry Planning can happen either in joint or operational space There is usually an infinity of possible desired trajectories How is the desired trajectory represented? Every point in time? Only start & final point? Via points? Movement Primitives x d = g t , α ( ) or θ d = f t , α ( )
Image of page 5

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Joint Space Planning What could one plan? Arbitrary trajectories from start to end Trapizoidal (or any aother kind of) velocity profiles Polynomials: 1.order: straight lines 2.order: parabolas 3.order: cubic splines 5 order: quintic splines Interesting: Analyze the shape of the trajectories in position, velocity, acceleration, and jerk space.
Image of page 6
Image of page 7
This is the end of the preview. Sign up to access the rest of the document.

{[ snackBarMessage ]}

What students are saying

  • Left Quote Icon

    As a current student on this bumpy collegiate pathway, I stumbled upon Course Hero, where I can find study resources for nearly all my courses, get online help from tutors 24/7, and even share my old projects, papers, and lecture notes with other students.

    Student Picture

    Kiran Temple University Fox School of Business ‘17, Course Hero Intern

  • Left Quote Icon

    I cannot even describe how much Course Hero helped me this summer. It’s truly become something I can always rely on and help me. In the end, I was not only able to survive summer classes, but I was able to thrive thanks to Course Hero.

    Student Picture

    Dana University of Pennsylvania ‘17, Course Hero Intern

  • Left Quote Icon

    The ability to access any university’s resources through Course Hero proved invaluable in my case. I was behind on Tulane coursework and actually used UCLA’s materials to help me move forward and get everything together on time.

    Student Picture

    Jill Tulane University ‘16, Course Hero Intern