Communications and Tracking
We are working on adapting the EM and Viterbi algorithms for use in various
communications and multitarget tracking problems.
If you'd like to discuss these topics, send mail to
Rick Perry or
Kevin Buckley.
-
Multiple User Maximum Likelihood Based Sequence Estimation for
Unknown, Frequency Selective, Fast Time Varying Channels
,
our paper from the November 2001
Asilomar
Conference on Signals, Systems, and Computers,
is available in
gzip'd
or zip'd postscript
and html formats.
Simulation code: download,
browse.
-
Direct and EM-based MAP Sequence Estimation
with Unknown Time-varying Channels
,
our paper from the May 2001
IEEE International Conference on Acoustics, Speech, and Signal Processing
(ICASSP-2001)
is available in
gzip'd or
zip'd postscript,
pdf,
and html formats.
-
MAP Sequence Estimation for Fading ISI Channels
,
our paper from the March 2001 Conference on
Information Sciences and Systems
(CISS),
at The Johns Hopkins University,
is available in
gzip'd or
zip'd
postscript and
html formats.
-
Time-Recursive Maximum Likelihood Based Sequence Estimation for
Unknown ISI Channels
,
our paper from the October 2000
Asilomar
Conference on Signals, Systems, and Computers,
is available in
gzip'd
or zip'd postscript
and html formats.
-
EM Algorithm for Sequence Estimation
over Gauss-Markov ISI Channels
,
our paper from the June 2000 International Conference on Communications
(ICC'2000),
New Orleans, Louisiana, is available in
gzip'd postscript and
html formats.
-
A New Pruning/Merging Algorithm for MHT Multitarget Tracking
,
our paper from the
May 2000 IEEE International Radar Conference
(RADAR 2000),
Alexandria, VA, is available in
gzip'd or
zip'd
postscript and
html formats.
-
Time-Recursive Number-of-Tracks Estimation for MHT
,
our paper from the April 2000 Signal and Data Processing of Small
Targets
(SPIE)
conference, Orlando, Florida, is available in
gzip'd postscript and
html formats.
-
Maximum Likelihood Source Localization using the EM Algorithm
to Incorporate Prior Parameter Distributions
,
our paper from the March 2000 IEEE Sensor
Array and Multichannel
(SAM) Signal Processing Workshop,
Cambridge, Massachusetts,
is available in
gzip'd postscript and
html formats.
-
Sequence Estimation over Linearly-Constrained Random Channels
,
our paper from the March 2000 Conference on
Information Sciences and Systems
(CISS), at Princeton University,
is available in
gzip'd postscript and
html formats.
-
EM Algorithms for Sequence Estimation over Random ISI Channels,
our paper from the October 1999 Asilomar converence,
is available in
gzip'd postscript and
html formats.
-
Trellis Structure Approach to Multitarget Tracking,
our paper from the March 1999
Adaptive Sensor Array Processing Workshop (ASAP'99),
is available in
gzip'd postscript and
html formats.
-
Multitarget List Viterbi Tracking Algorithm,
our paper from the November 1998 Asilomar conference,
is available in
gzip'd postscript and
html formats.
M simulation code is also
available for browsing
and download.
-
Old K-path L-best Viterbi simulation code for use with
Matlab or
Octave
is available for browsing or download in
gzip'd tar and
zip formats. The original
and older version in M
is also available for browsing
or download.
- Matlab plot showing detection of
dissimilar trajectories which have similar costs.
Artistically enhanced plot
which I use to explain multitarget tracking to my kids.
Internal reports:
- Dual Binomial EM Examples
M simulation code:
b1.m, b2.m,
fmin.m.
- Sequence Estimation EM Examples
M simulation code:
b3.m, b4.m,
vit.m, ber.m.
- Time-Recursive EM Examples
M simulation code:
b3r.m, b4r.m,
ber_r.m, ber_r3.m.
- Distributed EM Example
- EM ISI Examples,
and some simulations.
- EM ISI with time-varying channel coefficients
and some simulations.
- Some plots of time-varying channel coefficients.
- Kaleh-Vallet algorithm comparison plot,
from a paper in preparation (of interest only to the authors).
Notes about using the Viterbi algorithm with a Kalman
filter cost function for target tracking:
Some small problems about finding closed-form expressions
for the number of Viterbi states and permutations with
missed target events:
