| Author | Poramate Tarasak |
| Call Number | AIT Thesis no.TC-99-9 |
| Subject(s) | Wireless communication systems
|
| Note | A thesis submitted in partial fulfillment of the requirements for the degree of Master of Engineering. School of Advanced Technologies |
| Publisher | Asian Institute of Technology |
| Series Statement | Thesis ; no. TC-99-9 |
| Abstract | Trellis-coded modulation (TCM) is a well-known coding scheme that can improve the
performance over additive white Gaussian noise (AWGN) and frequency-flat fading channels
without bandwidth expansion. Block-coded modulation (BCM) has been studied as a
counterpart to TCM. TCM has been also studied over frequency-selective fading channels in
some research but very few have been concentrated on BCM. In this thesis, both BCM and
TCM schemes have been studied and compared over frequency-selective fading channels with
various kinds of equalizers. The equalizers considered are Viterbi equalizer (VE), soft-output
Viterbi equalizer (SOVE), sub-optimum soft-output equalizer (SSE) and Max-Log-MAP
equalizer. For TCM case, a combined equalizer-decoder has also been included. The channel
is modeled as two-ray equal power Rayleigh fading channel. Perfect channel estimation is
assumed. Both independent and correlated fading channels are considered.
Three TCM schemes are studied and compared with different types of equalizers. At a
bit error rate level of 10-5
, 7-8 dB gain over VE and SOVE is achieved with SSE and MaxLog-MAP equalizer for the 4-state nonparallel transition TCM and 8-state Ungerboeck's
TCM. Four-state Ungerboeck's TCM outperforms 4-state nonparallel TCM with VE and
SOVE, while the performance reverses with SSE and Max-Log-MAP equalizer.
Multistage decoding is compared with Viterbi decoding ofBCM. In one BCM scheme,
the results show that multistage decoding outperforms Viterbi decoding with VE and SOVE.
This may be due to the fact that multistage decoding can compensate some information that
has been lost from the hard-decision output of the equalizer.
A comparison of the performance between multistage decoding of BCM and Viterbi
decoding of TCM has been given. The simulation results show the comparable performance of
BCM to TCM in some system configurations. BCM has less decoding complexity.
A new analytical upper bound is derived and some approximations have been made to
evaluate the performance. The derived bound indicates the importance of the effective code
length (ECL) and the product of the ISI distance in the independent fading channel. For the
correlated fading channel, the important parameter are the ECL and the eigenvalues of a
covariance matrix, which is a function of maximum Doppler frequency, symbol duration, and
the ISI distance of the code.
This upper bound is applicable to a combined equalizer-decoder of TCM and separate
VE-decoder cases. The results show that the upper bounds are within 2-3 dB of the
simulations in the combined equalizer-decoder cases and within 1-5 dB in separate VEdecoder cases of both TCM and BCM. |
| Year | 1999 |
| Corresponding Series Added Entry | Asian Institute of Technology. Thesis ; no. TC-99-9 |
| Type | Thesis |
| School | School of Advanced Technologies (SAT) |
| Department | Department of Information and Communications Technologies (DICT) |
| Academic Program/FoS | Telecommunications (TC) |
| Chairperson(s) | Rajatheva, R.M.A.P.; |
| Examination Committee(s) | Erke, Tapio;Makelainen, Kimmo; |
| Scholarship Donor(s) | Asian Institute of Technology ;John F. Kennedy Foundation; |
| Degree | Thesis (M.Eng.) - Asian Institute of Technology |