Research On Equalization Technology In VLC System

In recent years, visible light communication based on LED has become the research focus in the field of communications for its distinct advantages including the abundance of unregulated bandwidth, high energy conversion efficiency, environmental friendliness and high security. In this thesis, the equalization technology in VLC is studied and the main work is listed as follows:1) Firstly, the thesis studies the optical channel of both directional and non-directional LOS links. The channel impulse response (CIR) of the directional LOS link is obtained by measured results of our research group’s VLC platform. Based on the simulation results of photon tracing model and the measured data, CIR of the non-directional link is obtained.2) Secondly, considering the time-invariant and low-pass characteristics of directional LOS channel, two pre-equalization schemes are proposed to extend the transmitted bandwidth, that is, analog and digital pre-equalization. It is shown by the measured data that the analog pre-equalization circuit can raise the 9MHz flat bandwidth (3dB) of the system to 26MHz, but there is a decrease in the amplitude of signal. When employing digital pre-equalization by an FIR filter of order 41 under the simulation conditions of this thesis, the 9MHz bandwidth can be raised to 74MHz. Furthermore, compared to post-equalization, pre-equalization shows better anti-noise performance.3) Thirdly, considering the time-variant and multipath characteristics of non-directional LOS channel, the adaptive equalization technique with training sequence is studied to combat the ISI at the receiver. The focus is on the LMS adaptive algorithm and the improved algorithms proposed by the predecessors. Given the fact that the parameters of versiera NLMS algorithm cannot optimize itself according to the optical channel, which changes slowly with the environment, an improved NLMS algorithm is proposed in this thesis. The improved NLMS algorithm uses variable step to speed up the convergence in the early iterations and to improve the adjustment accuracy, while in the later iterations, fix step is employed to make itself converge to a smaller steady-state error at the end of the training sequence. The simulation results show that the improved NLMS algorithm achieves better performance in different channel environments.4) Lastly, blind equalization is studied in non-directional LOS link which is not rather contaminated. The aim is to minimize the ISI at the receiver and simultaneously, to improve the bandwidth utilization mainly using the Bussgang-type blind equalization. Considering that decision-directed algorithm in poor channel environment cannot converge because of the unreliable current error, an improved decision-directed algorithm is proposed. The improved algorithm enhances the blind convergence ability by stopping the iteration with unreliable current error. The simulation results show that the steady-state mean square error is far smaller than the conventional one in a poor channel environment.