Due to increasing demand for bandwidth,spectrum congestion has become one severe technical challenge in radio frequency?RF?field.Visible light communication?VLC?attracts more attention because of its abundant bandwidth.Orthogonal Frequency Division Multiplexing?OFDM?is regarded as one of the promising technologies in VLC for its robustness to channel impairments,simple equalization,and efficient implementation using the fast Fourier transform?FFT?.Unlike conventional RF systems,VLC is based on intensity modulation and direct detection?IM/DD?.Because the light intensity is modulated with data,non-negative signal should be transmitted through channel.Asymmetrical-Clipped Optical OFDM?ACO-OFDM?is a kind of light power efficient OFDM technology.In the thesis,channel equalization algorithms are investigated to improve performance for ACO-OFDM systems.The main works are as follows:1.Firstly,tailored to an ACO-OFDM VLC system,a constant norm algorithm?CNA?blind equalization algorithm is proposed.We analyze the parameters of initial value and step size of CNA,in order to optimize them for faster convergence and better equalization performance.At the receiver,compared to conventional equalization algorithm,the proposed algorithm doesn't need training sequence?TS?and pilots,which improves the light spectrum efficiency and effective data rate.According to the numerical results,the ACO-OFDM signal equalized by the CNA blind equalization algorithm has better mean square error?MSE?than that equalized by constant modulus algorithm?CMA?.In addition,bit error rate?BER?performance of CNA,CMA blind equalization and TS-based equalization is measured in ACO-OFDM VLC system.The numerical results show that in a VLC system operating at the fixed data rate,the BER performance of CNA blind equalization algorithm is better than that of CMA.2.Secondly,we propose an algorithm based on K-means clustering for ACO-OFDM VLC systems.The algorithm exploits the constellation performance to categorize the received signal.Since K-means doesn't need TS,the system spectrum efficiency is effectively improved.Numerical results show that,the nonlinear impairments on ACO-OFDM signal in VLC systems are mitigated by the proposed K-means clustering algorithm.Thus,it improves the performance of VLC systems.At the BER threshold of hard-decision forward error correction?HD-FEC?,3.8×10-3,the K-means clustering algorithm provides 5dB SNR gains compared to without the K-means clustering algorithm. |