|
Course Info
Handouts
Problem
Sets
Download
|
Announcements:
- Zip file for
Project 1 available under downloads (3/2/05)
|
Course
Information:
| Instructor: |
Derek Rowell
Room 3-142
x 3-6206
drowell@mit.edu
|
| TA: |
Brian
Anthony (banthony@mit.edu) |
| Location: |
Room
1-246 |
| Time: |
MW
11:00am - 12:30pm |
| Credit: |
Graduate H
Level
Satisfies MSME distribution requirement in System
Dynamics and Control
|
This course
will provide a solid theoretical foundation for the analysis and processing
of
experimental data, and real-time experimental control methods. The course
will emphasize
practical problems with laboratory exercises. The course is designated
as a d'Arbeloff
Laboratory 'gateway' subject - that is, it is considered to be an essential
lead-in to more
specialized courses in information technology.
Topics covered:
1) Introduction
to Real-Time Computation: Data converters (A/D, D/A), machine architecture,
software considerations.
2) Review of Linear Continuous-Time Signal Processing: Fourier
methods, Laplace transform, convolution, frequency/time domain processing.
3) Discrete-Time Signal Processing: The z transform, difference
equations, relationship between F(z) and F*(jw), mappings between s-domain
and z-domain, inverse z transform. Discrete-time stability.
4) Sampling and Reconstruction: Sampling theorem, aliasing, quantization,
sampled data systems, cardinal (Whitaker) reconstruction, zero-, first-,
second-order hold reconstructors, interpolators, non-resetting reconstructors,
matched filtering, etc.
5) Discrete Spectral Analysis: The DFT and its relationship to
the continuous FT, the FFT and implementations (decimation in time and
frequency), radix-2 implementation, leakage, windowing. Uses of the
DFT: convolution - (overlap & add, select savings) correlation.
Random processes, power spectral density (PSD) estimation - methods
of smoothing the periodogram (Welch's method, windowing the correlation
function, etc). System identification.
6) Real-Time Simulation Methods using Difference Equations: Impulse-,
step-, ramp-invariant simulations. Tustin's method, matched poles/zeros,
bilinear transform methods.
7) Filter Design - Continuous and Discrete: Butterworth, elliptic,
Tchebychev low-pass filters. Low-pass design methods based on continuous
prototypes. Realizations. Conversion to high-pass, band-pass, band-stop
filters.
8) Introduction to System Identification
9) Introduction to Discrete-Time Feedback
The course
will have a significant hands-on component with laboratory exercises
requiring
analysis/processing of data captured from experimental apparatus.
|
