Full text
| |
| Author(s): |
Guilherme Silveira Rabelo
Total Authors: 1
|
| Document type: | Master's Dissertation |
| Press: | Campinas, SP. |
| Institution: | Universidade Estadual de Campinas (UNICAMP). Faculdade de Engenharia Elétrica e de Computação |
| Defense date: | 2010-12-04 |
| Examining board members: |
Daniel Benevides da Costa;
Renato da Rocha Lopes
|
| Advisor: | Michel Daoud Yacoub |
| Abstract | |
This work evaluates the performance of a wireless link under severe fading conditions. To achieve this, a new probability distribution, the ?-µ Extreme, is investigated as a fading model to describe the propagated signal under such conditions. Adjustments to field data published elsewhere in the literature, collected in an enclosed environment, and other fading data collected by the author's team show that the ?-µ Extreme distribution is flexible and suitable to model these scenarios. Link performance measures are then obtained for non-coherent modulation schemes, where it is shown that, due to an intrinsic characteristic of the ?-µ Extreme fading channel, the bit error rate is lower-bounded by a non-nil minimum value. In the same way, for anyone of the M-branch basic diversity combining techniques, namely selection combining, maximal ratio combining, and equal gain combining, the bit error rate is still lower-bounded, although the minimum achievable value decreases exponentially with the increase in the number of diversity branches. Formulations for the outage probability as well as mean signal-to-noise ratio at the output of the combiners are also derived. Interestingly, it is shown that in extreme fading conditions, the selection combiner may outperform the equal gain combiner. In addition, this works also derives a new closed-form series-type formulation for the multivariate a-µ joint distribution, which includes as particular cases the largely used fading distributions Weibull, Nakagami-m, and Rayleigh. Written in terms of functions already available in literature, the formulation is easily computable and very general for it assumes arbitrary correlation scheme and parameters. Even though the new formulation is given in terms of an infinite series, it is shown that a truncated version is still a valid distribution. Finally, in order to demonstrate the applicability of the new results, the envelope cumulative distribution at the output of the selection combiner is evaluated using the approximate and exact approaches. The results show that that only a few terms is sufficient to provide an excellent fit between approximate and exact results (AU) | |
| FAPESP's process: | 09/04291-4 - Wireless Systems Performance Under Severe Fading Conditions |
| Grantee: | Guilherme Silveira Rabelo |
| Support Opportunities: | Scholarships in Brazil - Master |