\( \newcommand{\kgoal}[1]{\mathord{\class{chip chip-goal}{\kern0.14em {#1}\kern0.14em}}} \newcommand{\kspring}[1]{\mathord{\class{chip chip-spring}{\kern0.14em {#1}\kern0.14em}}} \newcommand{\kforce}[1]{\mathord{\class{chip chip-forcing}{\kern0.14em {#1}\kern0.14em}}} \newcommand{\kphase}[1]{\mathord{\class{chip chip-phase}{\kern0.14em {#1}\kern0.14em}}} \newcommand{\kbasis}[1]{\mathord{\class{chip chip-basis}{\kern0.14em {#1}\kern0.14em}}} \newcommand{\kdemo}[1]{\mathord{\class{chip chip-demo}{\kern0.14em {#1}\kern0.14em}}} \newcommand{\krepro}[1]{\mathord{\class{chip chip-repro}{\kern0.14em {#1}\kern0.14em}}} \)

DMP 101

A friendly, intuition-first guide to dynamic movement primitives

Author
Affiliations

Fares J. Abu-Dakka

Assistant Professor of Mechanical Engineering

Robot Learning & Control Lab (REACH Lab), New York University Abu Dhabi

Welcome

Imagine showing a robot how to wave just once, and then watching it wave again somewhere new, quicker or slower, larger or smaller, without ever needing a second lesson. That small bit of magic is what a dynamic movement primitive (DMP) makes possible, and along the way through this book, you will discover how it works.

Bring your Python and your curiosity; you can join us even if you have never taken a control course. We will meet each new word as it comes along, starting every chapter beside a robot and exploring every idea in a notebook you can run and break.

As you explore, your notebook work stays in your browser on this computer; the server does not keep it.
Download your notebooks before you clear your browser data or switch computers, or you will lose your changes.

The three notebooks

  • 01 · The discrete playground: begin with a motion you load or draw, and once you have fitted it, explore what happens when you move the goal and the start or change the speed.
  • 02 · The periodic playground: teach the robot a rhythm, and then watch how the beat changes with its frequency while the swing grows or shrinks with its size.
  • 03 · Create your own: give the robot something of your own to do, from writing your initials to waving, and along the way, break a DMP on purpose.

How to read this book

As we walk through each chapter, the same four kinds of box will keep you company:

  • Intuition gives you a picture to hold in your head before the symbols arrive.
  • The math puts that picture into equations, with each one followed by a sentence that reads it aloud.
  • Try it takes you to a notebook and the exact cell to open, with a little guidance on what to look for.
  • Gotcha walks you through the chapter’s one trap, showing what goes wrong, why it happens, and how to find your way out.

To help you find your bearings in equations, each idea keeps one color everywhere: the goal \(\kgoal{g}\), the spring–damper term \(\kspring{\alpha_z(\beta_z(g - y) - z)}\), the forcing term \(\kforce{f(x)}\), the phase \(\kphase{x}\), the basis functions \(\kbasis{\Psi_i}\), the demonstration \(\kdemo{y_d}\) and the reproduction \(\krepro{y}\). When we move from equations to plots, a second cue, such as a dashed line for the demonstration, helps you follow the story, so you never have to rely on color alone.

The model equations you will meet come from one survey paper and three older ones, all listed on the Sources page, so you can follow their numbers back to where the story began.