The Digital Signal Processing (DSP)
Laboratory at Florida International University was established in August
1994, with the two-fold purpose of developing research work in various areas of DSP, while providing our graduate and undergraduate students with
an opportunity to experience DSP research and learn the
specific skills sought by industrial employers in this field.
Accordingly, there have been two major
thrusts in our activities:
1 |
Real-time implementation of Digital Processing
Algorithms in specialized DSP chips. |
2 |
Development of innovative DSP
algorithms for the processing of specific signals, or for purpose-specific applications. |
The members of the laboratory acquire skills
in the real-time implementation of DSP algorithms using
specialized DSP chips through the development of assignments
and projects. The "Real-Time DSP Implementations"
course in our curriculum fosters the development of these skills.
The DSP Laboratory has developed and implemented algorithms
for the processing of signals in diverse contexts. Several of our projects have involved
processing signals from biomedical transducers, such as Blood Volume Pulse
photoplethysmographs, and Electroencephalogram (EEG) and Electromyogram (EMG) electrodes.
We have applied processing to the signals from these transducers towards the development
of alternate human-computer interfaces for the benefit of users with severe motor
disabilities. Recently, we have worked in integrating an infrared video eye gaze tracking
system with the EMG-driven interface.
The DSP Lab has carried out research in the area of
Deconvolution of the Doppler-Azimuth Radar Spectrum, by which the spatio-temporal data
obtained from a multi-element radar array is enhanced to obtain a more defined
characterization of objects of interest, while minimizing the effects of unwanted
components.
The DSP Lab has also been involved in applied research for
the local manufacturing industry. We have developed a self-tuning motion control system
for a local Computerized Numerical Control (CNC) company. In this project, a DSP chip executes an algorithm to obtain a dynamic model of a CNC
machine (such as a lathe or milling machine). It is then possible to implement a real-time
inverse controller to compensate for the dynamic characteristics of the plant, yielding an
improved performance.
|