Instructors: Chris Atkeson,
cga at cmu
TA: Allison Funkhouser, afunkhou at andrew
Time: TR 3-4:20PM
Place: NSH 3002
Events of interest
Aaron Dollar, Tuesday, March 31st 1 PM GHC 6115
Mechanical Intelligence in Robotic Manipulation: Towards Human-level
Dexterity in Robotic and Prosthetic Hands
The human hand is the pinnacle of dexterity - it has the ability to
powerfully grasp a wide range of object sizes and shapes as well as
delicately manipulate objects held within the fingertips. Current
robotic and prosthetic systems, however, have only a fraction of that
manual dexterity. My group attempts to address this gap in two main
ways: the mechanical design of effective hands and the study of human
hand function and use as inspiration and performance benchmarking. In
terms of hand design, we strongly prioritize passive mechanics,
including incorporating adaptive underactuated transmissions and
carefully tuned compliance, and seek to maximize open-loop
performance while minimizing complexity. To motivate and benchmark
our efforts, we are examining human hand usage during daily
activities as well as quantifying functional aspects such as
precision manipulation workspaces. Besides describing these efforts,
I will touch on other work in the lab related to legged robots, novel
fabrication techniques, modular robots, and the study of non-human
Ex Machina (movie)
opens April 10.
For more information, take a look at
version of this course.
Jan 13: Introduction
Jan 15: Intro to Vision
Jan 20: 2D Debris Task: Registration, Hands, Choosing a target,
Jan 22: Trajectories
Jan 27: Forward and Inverse Kinematics
Jan 27: Using Matlab
A basic tutorial for vectors and matrices,
Jan 27: Inverse Kinematics Using Function Optimization
Robotics: redundant inverse kinematics.
Using Matlab's fminsearch and fminunc.
Using Matlab's fminsearch and fminunc, with
Using Matlab's fmincon.
Jan 29 - Feb 3: 3D Orientation
Feb 5: 3D Vision
Finding planes: Hough transform,
Principal component analysis.
Feb 5: 3D Debris Task
Matlab code for DRC robot forward kinematics.
Feb 10: Collision Detection
Feb 10-12: Manipulation: Planning
Straight line paths in joint space and Cartesian space,
Rapidly Exploring Random Trees (RRT),
Feb 17: Actuation: Force/torque sources, position sources with compliance
Feb 17: Feedback Control
Feb 19: Jacobian
automating derivation of
two link arm Jacobian
Feb 19: Dynamics
gravity compensation, full dynamics,
Feb 24: Simulation
Feb 24: Model Building
Kinematic calibration, Mass calibration, Dynamic calibration.
Feb 24: Learning Models
k-nearest neighbors algorithm,
Kernel regression and locally weighted regression.
Neural net 2,
Neural net learning,
Feb 26: Learning From Practice:
Feb 26: Locomotion: legged balance
Feb 26: Locomotion: running
March 3: Locomotion: walking:
static: hexapod, quadruped, biped,
passive dynamic walking.
March 3: Locomotion: walking and running on rough terrain
Footstep planning, A* search, Trajectory optimization.
March 5: Project help session (Allison)
March 9-13: No Class
March 17: Hands:
Utah-MIT hand 2,
March 19: HRI, Assignment 4, humanoid demo:
March 24: Hands II
March 26 Internal Sensors:
spindle, Golgi Tendon Organ
March 31: Skin, and tactile sensing:
March 31: Robot Touch/Force Sensors: skin, finger, wrist, joint, ankle, base.
GT fabric skin
400 receptors, 1 trillion smells
April 2: Project help session (Allison)
April 7: Pattern Learning:
April 7: Clustering
April 9: Learning Probabilities
April 9: Reinforcement Learning
April 14: Artificial People
April 14: Ethics, Policy Issues, Robots in entertainment
April 16: No Class
April 21: Learning From Demonstration
Inverse optimal control.
April 23: ?
April 28: Project Presentations
April 30: Project Presentations
May 10: Project Writeups Due
Ways to capture movies of your assignments
Assignment 0: Due Jan 19.
Send email to Chris and Allison: Who are you? Done any robotics?
Any project ideas? Google and send me some interesting URLs.
Be sure your name is obvious in the email, and you mention the course
name or number. We teach more than one course, and a random email from
firstname.lastname@example.org is hard for us to process.
Assignment 1: Due Jan 20.
Watch a movie or TV show or read a book that has a significant humanoid
robot character. Be prepared to discuss this character in class.
Current movies: Big Hero 6 and Interstellar.
Assignment 2: Inverse Kinematics: Due Feb 11.
Assignment 3: Vision: Due Mar 4.
Assignment 4: Animation: Due ?.
* means with high probability it could turn into a summer job.
* Build a soft robot and control it. See my web pages on soft robots
* Build Baymax
* Build a "robot theater/diorama" for STEAM education.
* Help me build a sensor network for the Sarcos humanoid.
* Build a prototype robot skin (with CGA help).
* Figure out how to make the Robotis-Mini (Darwin) useful to the course.
Get a microprocessor to do something involving the real world (sensing
or actuation). *Get an Intel Edison working.
Simulate something 3D in Matlab (hopping, walking, grasping, ...)
Simulate something 3D in ODE
Simulate something 3D in Gazebo
Develop a vision system to recognize your friends (Use a Kinect).
Develop a speech system for robot commands or robot interaction.
Train a neural net or other learning approach to do something interesting.
Implement some aspect of robotics that we discussed, or something
that interests you.
Survey the state of the art in some aspect of robotics.
Research some topic of interest to you.
Things to Think About
Info on Humanoids
Other Humanoids Courses (mostly graduate) and Research Groups
Course Poster, suitable for framing.