16-743 ROBOT CONTROL (Spring 1995)

Lectures:

TTh 1:30 - 2:50
Hamburg Hall 1510

Instructor:

Yangsheng Xu
xu+@cs.cmu.edu
WeH 1319, x8-3737

Course Outline:

The course is designed to introduce some fundamental methodologies of designing feedforward and feedback control schemes for robot manipulators, and discuss some basic control problems in the area of robotics.

• We will review some basic concepts in geometry, kinematics, and discuss extensively on manipulability measure and ellipsoid.
• We will introduce redundancy control of the robot manipulators and local and global approaches to the problems.
• We will study the coordination control issues for multiple arm motion and force control.
• We will discuss the basic theories and properties of manipulator dynamics, actuator dynamics, and the use of dynamic model.
• We will introduce computed-torque method and its approximate versions, and the classical joint controller.
• We will introduce optimal control for outerloop with LQ method, and robust control with saturation-type approach.
• We will formulate adaptive controllers for model uncertainty, and discuss some fundamental issues such as stability and persistence of excitation.
• We will discuss the problem of stiffness control and hybrid position force control.

Schedule:

1/17, Class overview, robot geometry
1/19, Kinematics review
1/24, Manipulability ellipsoid and measure
1/26, Dynamic manipulability
1/31, Kinematic redundancy
2/2, Local approach to redundancy control
2/7, Global approach to redundancy control
2/9, Coordination control: dynamics issues
2/14, Coordination control: control issues
2/16, Dynamics, force, inertia, and energy
2/21, Manipulator dynamics equation
2/23, Structure and properties of manipulator dynamics
2/28, Parameter linearization and energy passivity
3/2, Midterm exam
3/7, State space and Cartesian space representation
3/9, Feedback linearization and actuator dynamics
3/14, Computed-torque method: feedforward
3/16, Computed-torque method: feedback
3/28, Approximate Computed-torque method
3/30, Classical joint control
4/4, Optimal outerloop design
4/6, Robust control
4/11, More on robust control
4/13, Adaptive computed-torque method
4/18, Inertia-based adaptive control
4/20, Passivity-based adaptive control
4/25, Composite adaptive control
4/27, Stiffness control
5/2, Hybrid position/force control
5/4, Class review

Grading:

Grades will be determined by performance in the following 3 areas:

• Problem sets: 30%
• Project: 30%
• Midterm exam: 20%
• Final exam: 20%

Problem Sets:
Every problem set has equal weight. Each problem set is worth full credit when turned in at the beginning of class on the due date. It is worth half credit after that.

Project:
A list of projects involving experimental and simulation studies will be given around early March. Students can select project based on his/her interest. Project assignment will be executed by only one person independently. The due date for the assignment will be given when the project is assigned. Under certain circumstance, students also can propose a project and execute it with an approval of the instructor. A ten or more page description are due on the date specified for full credit.

Exam:
Two oral exams will be given at the mid and the end of class. The midterm will cover all preceding material, and the final exam will cover post-midterm material.

Reference Books:

• "Control of Robot Manipulators"
  Lewis, Abdallah, and Dawson
  Macmillan Publishing, 1993
• "Advanced Robotics: Redundancy and Optimization"
  Yoshihiko Nakamura
  Addison-Wesley Publishing Company, 1991

Teaching Assistant:

Chris Lee
WeH 1306, x8-2429
chrislee@cs.cmu.edu