Light-emitting diode (LED) luminaires are ubiquitous and widely used in our daily life. Because of energy efficiency and illumination requirements, LED lamps are required to dim. In addition to the lighting function, visible light communication (VLC) is being considered as a potential technology beyond 5G. Orthogonal frequency-division multiplexing (OFDM) is a popular modulation choice and most recently spatial optical OFDM (SO-OFDM) has been used with multiple LEDs. SO-OFDM is a low complexity way of constructing OFDM signals by sending a series of narrow-band emissions from different LEDs and allowing them to sum in space. The integration of dimming within an SO-OFDM framework remains an open area of research. In this thesis, an optimal dimming strategy for optical power allocation is
proposed based on SO-OFDM in visible light communication systems under different
illumination requirements, i.e., dimming. Firstly, a non-convex problem is formulated to maximize the capacity of SO-OFDM under a specified dimming target. Furthermore, this non-convex problem is broken into a group of convex problems through approximations and constraint relaxation.
The proposed scheme can optimize the power allocation in SO-OFDM and achieve the dimming target at the same time. According to the proposed power allocation scheme, more of the power is allocated to subcarriers with low noise variance. This phenomenon makes sense to take advantage of good channels when power is limited and is similar to the conventional water-filling method. When the dimming target and signal-to-noise ratio (SNR) is sufficiently high,
the optimal distribution approaches a uniform distribution that agrees well with conventional water filling in a high SNR regime.
Compared with traditional digital dimming, the proposed spatial dimming method can not only use spatial domain for dimming but also allocates power to different subcarriers according to channel conditions which is more efficient.
Analytical and numerical results imply that the proposed scheme converges quickly and performs well compared with traditional power allocation methods. / Thesis / Master of Applied Science (MASc)
| Identifer | oai:union.ndltd.org:mcmaster.ca/oai:macsphere.mcmaster.ca:11375/26689 |
| Date | January 2021 |
| Creators | Donghan, Li |
| Contributors | Steve, Hranilovic, Electrical and Computer Engineering |
| Source Sets | McMaster University |
| Language | English |
| Detected Language | English |
| Type | Thesis |
Page generated in 0.0026 seconds