Teaching
The mechanics of robots, comprising kinematics, dynamics, and
trajectories. Planar, spherical, and spatial transformations and
displacements. Representing orientation: Euler angles, angle-axis, and
quaternions. Velocity and acceleration: the Jacobian and screw
theory. Inverse kinematics: solvability and singularities. Trajectory
planning: joint interpolation and Cartesian trajectories. Statics of
serial chain mechanisms. Inertial parameters, Newton-Euler equations,
D'Alembert's principle. Recursive forward and inverse dynamics.
This course introduces modern probabilistic approaches towards
creating intelligent systems, where rationale decision-making is
phrased in terms of maximizing expected utility. Basic concepts of
search are introduced, leading to search under uncertainty, Markov
decision processes, Bellman's equations, and reinforcement learning,
Bayes nets are introduced to reduce dependencies among
variables. Hidden Markov models and partially observable Markov
decision processes are introduced to handle uncertainties in state.
Modeling and identification of the mechanical properties of robots and
their environments. Review of probability and statistics. Parametric
versus nonparametric estimation. Linear least squares parameter
estimation, total least squares, and Kalman filters. Nonlinear
estimation and extended Kalman filters. State estimation. Specific
identification methods for kinematic calibration, inertial parameter
estimation, and joint friction modeling.
The Robotics Seminar is required for all new robotics students in the
fall and spring terms of the first year. It can also be taken by
students wishing to learn more about robotics at Utah. It is a chance
for the Utah robotics community to get together on a weekly basis. The
course features technical presentations by robotics graduate students,
faculty or distinguished visitors, and discussions of recent trends
and important results elsewhere.