الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 91 |  |  |
| | | from | 91 | | |
Abstract
Communication is essential in our lives. The sector of communications has grown
rapidly as technology has advanced and its needs have increased. Signals that were initially
transmitted in the analogue domain are increasingly being transmitted in the digital domain.
To improve transmission, single carrier waves are being replaced by multi-carrier waves.
Multi-carrier system such as Orthogonal Frequency Division Multiplexing (OFDM) is
increasingly being used. The data is carried from the transmitter end to the receiver end
using orthogonally arranged subcarriers in the OFDM system. The addition of a guard band
in this system resolves Inter Symbol Interference (ISI), while the increased number of
subcarriers minimizes noise. OFDM is used extensively in Visible Light Communication
(VLC) due to its ability to support high speed transmission. However, the high Peak to
Average Power Ratio (PAPR) of OFDM systems has been one of their major drawbacks.
The LED’s linear region might be surpassed if the ratio of peak power to average power is
too high. The result is a degradation in performance due to non-linear distortion, which
changes the superposition of the signal spectrum.
In this thesis, there are two different proposed techniques to deal with the problem
of PAPR in OFDM system. The first proposed technique makes use of wavelet and
companding with each other for denoising the channel effect and reducing the PAPR value.
The second proposed technique utilizes deep learning algorithms for self-learning the
appropriate weights that reduces Bit Error Rate (BER) and PAPR losses. Both techniques have achieved improvement according to BER and PAPR in comparison with the traditional PAPR reduction techniques.