Phase linear regression model with adaptive modulation technique for filtered-orthogonal frequency division multiplexing
One of the applications of Orthogonal Frequency Division Multiplexing (OFDM) is teleconferencing as it has shown to have ten times better downlink response time compared to 3G. Despite the robustness against multi-path fading and ease of OFDM implementation, OFDM still suffers from high sidelobes in...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2021
|
| Subjects: | |
| Online Access: | http://eprints.utm.my/106983/1/AkramMuhammadRazeePFTIR2021.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | One of the applications of Orthogonal Frequency Division Multiplexing (OFDM) is teleconferencing as it has shown to have ten times better downlink response time compared to 3G. Despite the robustness against multi-path fading and ease of OFDM implementation, OFDM still suffers from high sidelobes in frequency domain and high Bit Error Rate (BER). In the OFDM system, frequency mask regulation is adopted to overcome the issues, whereby a portion from both ends of the bandwidth is reserved as a frequency guard band. This thus degrades the spectral efficiency and does not reduce the Out of Band Emission. Therefore, the goals of this research are to investigate the impact of filter types, and design a new filter in the filtered-OFDM (f-OFDM) system. Initially, several Infinite Impulse Response (IIR): Butterworth, Chebyshev, Elliptic; and Finite Impulse Response (FIR): Equiripple, Bohman, and Hamming filters were evaluated in terms of magnitude response, phase response and group delay. The results showed that Elliptic and Butterworth achieved good performance in magnitude response and phase response for FIR and IIR, respectively. The Equirriple was applied by varying the filter order in f-OFDM under different modulation schemes and channel models. It was found that f-OFDM achieved BER 10−6 at 11.9 dB while OFDM 12.4 dB, particularly when Equiripple with 512th order was used. However, the complexity and group delay also increased. By using the results of the BER performance for the optimum Equiripple design in the f-OFDM, the link adaptation model was proposed. Next, because IIR is a non-linear phase filter, a new method called Phase Linear Regression Model, a combination of mean square error and best fit was proposed by considering the magnitude and phase responses in designing Butterworth for the f-OFDM waveform. The new IIR design with 5th order achieved BER 6.667 × 10−7 at SNR 12 dB, while Equirriple filter 1.552 × 10−5 at the same SNR and filter order for the Quadrature Phase Shift Keying modulation. In conclusion, the feasible type of IIR (Butterworth) and FIR filter (Equiripple) can improve the BER performance for the f-OFDM system compared with OFDM. The findings on the link adaption model as well as the new IIR filter should be further investigated for the multi-cell f-OFDM system |
|---|