Capacity Analysis And Constellation Design In Visible Light Communication Systems

With the rapid development of wireless transmission requirements, the wireless spectrum resources are becoming more and more limited. Although many new techniques have been proposed to improve the utilization of wireless spectrum resources, the spectrum scarcity prob-lems have not been solved yet. Recently, as a supplementary technique for radio frequency (RF) communications, visible light communications (VLC) has attracted considerable interest in both academy and industry, which is becoming one of the most hot topics nowadays. The VLC has some characteristics that are different from the RF communications, and thus the developed theory and analysis in RF communications are not directly applicable to VLC. Hence, the key problems of the VLC system are studied in this dissertation. The main work is composed of the following four parts:Chapter 2 derives the theoretical expressions of the upper and lower bounds on the channel capacity for the point-to-point VLC system. Under the input-independent Gaussian noise, this chapter considers the nonnegativity, peak optical intensity and average optical intensity con-straints. The lower bound on the capacity is obtained based on the entropy power inequality and the calculus of variations, while the derivation of the upper bound on the capacity is carried out based on the relative entropy. In addition, the performance gap between the derived upper and the lower bounds is analyzed. Numerical results show that the presented bounds are very tight, which verifies the accuracy of the derived bounds.In Chapter 3, the capacity bound for the multiple input multiple output (MIMO) VLC sys-tem is derived, the positions of the receivers are planned as well. In this chapter, assuming that perfect channel state information is known at both the transmitter and the receiver, the MI-MO channel is converted into the parallel, non-interfering channels by using the singular value decomposition. With the nonnegativity, peak optical intensity and average optical intensity con-straints, the theoretical expression of the lower capacity bound is derived by using the principle of the entropy power inequality and the calculus of variations. Based on the lower capacity bound, the optimization problem for planning the receivers is proposed, and a particle swar-m optimization based iteration algorithm is employed to solve the problem. Numerical results verify the accuracy of the lower capacity bound and the receiver deployment scheme.Chapter 4 designs the signal constellation of the indoor VLC system with input-independent Gaussian noise. In this chapter, a point-to-point indoor VLC system using pulse amplitude mod-ulation (PAM) is considered, where the noise is assumed as the input-independent Gaussian noise. Under the constraints of the nonnegativity, peak optical intensity and average optical intensity, a joint constellation design scheme is proposed based on two existing constellation design schemes. In the proposed scheme, the PAM constellation is characterized by three pa-rameters, i.e., the intensities of the constellation points, the probabilities of the constellation points, and the modulation order. After that, the optimization problem for designing the signal constellation is presented, and an improved Blahut-Arimoto iteration algorithm is proposed to solve the problem. To reduce the computational complexity, a sub-optimal algorithm with low complexity is also given. Numerical results verify the performance of the proposed algorithms.Chapter 5 focuses on the constellation design for the outdoor VLC systems. In this chap-ter, the channel gain is determined by the atmospheric attenuation, atmospheric turbulence and pointing errors, which is a constant over a coherence interval. Under the non-negativity, peak optical intensity and average optical intensity constraints, the signal space is constructed. For the coherent and non-coherent detections, the corresponding distance criterions are provided, re-spectively. Based on the distance criterions, the optimization problem for constellation design is formulated, and then an interior point method based algorithm is given to solve the problem. To evaluate the system performance, the average constellation error probabilities are also derived. Numerical results are presented to show the performance of the derived signal constellation.