Introduction: Anatomy of a

Manipulation System

MIT 6.4210/2: Robotic Manipulation

Fall 2023, Lecture 1

Follow live at https://slides.com/d/sE8qYNg/live​

(or later at https://slides.com/russtedrake/fall23-lec01)

Course Info

Website: http://manipulation.mit.edu

6.4210 is a CI-M (Communication Intensive in the Major)

Communications Instructors from Comparative Media Studies/Writing

Recitations Fridays afternoons (starting Sept 15)

Logistics

  • Students: Make sure you're on Piazza
    • CI-M (6.4210) material on Canvas.
  • Review the course guidelines.
  • Read/comment on the lecture notes

 

  • ~ weekly problem sets throughout the class, due Wed.
  • Final project.

from the course website:

Today

  • What do I mean by "manipulation"? (with examples!)
  • The systems theory perspective
  • Anatomy of a manipulation system
  • Goals for the course

Matt Mason says

  • Definition 1 (etymological). Manipulation refers to the activities performed by hands.
  • ...
  • Definition 5. Manipulation refers to an agent’s control of its environment through selective contact.

Matthew T. Mason.  Toward Robotic Manipulation.  Annual Review of Control, Robotics, and Autonomous Systems, 1:1-28, 2018.

Autonomous "Open-world" Manipulation

  • Definition 5. Manipulation refers to an agent’s control of its environment through selective contact.

 

  • Even broader: "Open-world" manipulation requires:
    • rich perceptual understanding of the environment,
    • "common-sense" understanding of objects,
    • the ability to make long-term (task-level) plans, and combine them with fine (joint-level) motions.

Motivation

Consider the task of loading a dishwasher...

 

(One project I've worked on at TRI Robotics)

Low-level feedback control from cameras / sensors

(continuous time/state/action)

Scene-understanding +

Task-level Planning

(discrete/symbolic)

The world is even more "open" when the manipulation system goes mobile.

One goal for this year:

more mobile manipulation

(in lecture + projects)

Manipulation from the dynamics and controls perspective...

To be clear: we're not just controlling the arm

state of the robot x state of the environment

Levine*, Finn*, Darrel, Abbeel, JMLR 2016 

Visuomotor policies

perception network

(often pre-trained)

policy network

other robot sensors

learned state representation

actions

Manipulation is much more than "pick and place"!

The anatomy of a modern manipulation system

What is Drake?

drake.mit.edu

Also:

https://github.com/RobotLocomotion/drake-ros

https://github.com/RobotLocomotion/drake-external-examples

...

Logistics

  • Students: Make sure you're on Piazza
    • CI-M (6.4210) material on Canvas.
  • Review the course guidelines.
  • Read/comment on the lecture notes

 

  • ~ weekly problem sets throughout the class, due Wed.
  • Final project.