| 摘要: |
| 在室内可见光通信中符号间干扰和噪声会严重影响系统性能，K均值(K-means)均衡方法可以抑制光无线信道的影响，但其复杂度较高，且在聚类边界处易出现误判。提出了改进聚类中心点的K-means(Improved Center K-means，IC-Kmeans)算法，通过随机生成足够长的训练序列，然后将训练序列每一簇的均值作为K-means聚类中心，避免了传统K-means反复迭代寻找聚类中心。进一步，提出了基于神经网络的IC-Kmeans(Neural Network Based IC-Kmeans，NNIC-Kmeans)算法，使用反向传播神经网络将接收端二维数据映射至三维空间，以增加不同簇之间混合数据的距离，提高了分类准确性。蒙特卡罗误码率仿真表明，IC-Kmeans均衡和传统K-means算法的误码率性能相当，但可以显著降低复杂度，特别是在信噪比较小时。同时，在室内多径信道模型下，与IC-Kmeans和传统Kmeans均衡相比，NNIC-Kmeans均衡的光正交频分复用系统误码率性能最好。 |
| 关键词: 可见光通信 光正交频分复用 多径信道 信道均衡 K-means算法 反向传播神经网络 |
| DOI：10.20079/j.issn.1001-893x.231031003 |
|
| 基金项目:国家自然科学基金资助项目(61875080)；甘肃省科技计划(22JR5RA276，22JR5RA274，23YFGA0062，2022A-215) ；兰州理工大学博士科研启动经费(061903) |
|
| Channel Equalization Based on Improved K-means Algorithmfor O-OFDM Indoor Visible Light Communication |
| JIA Kejun,LIAN Jianglong,ZHANG Changrui,LIN Ying |
| (School of Computer and Communication,Lanzhou University of Technology,Lanzhou 730050,China) |
| Abstract: |
| Indoor visible light communication(VLC) is susceptible to inter-symbol interference(ISI) and background noise,which can severely affect system performance.Equalization based on K-means algorithm can effectively attenuate the ISI problem.However,it has high complexity issue and unsatisfactory equalization effect.The authors propose two improved methods to address these issues.In response to the high complexity of the K-means clustering algorithm and the issue of misclassification at cluster boundaries,an improved K-means algorithm called Improved Center K-means(IC-Kmeans) is proposed.This method involves the generation of sufficiently long training sequences,with the mean of each cluster in the training sequence serving as the K-means cluster center,which eliminates the need for the traditional iterative search for cluster centers in K-means.Furthermore,a neural network-based IC-Kmeans algorithm(NNIC-Kmeans) is introduced.NNIC-Kmeans utilizes a back propagation(BP) neural network to map the received two-dimensional data into a three-dimensional space,thereby increasing the separation between mixed data from different clusters and improving classification accuracy.Monte Carlo simulations of bit error rate(BER) demonstrate that IC-Kmeans and traditional K-means exhibit comparable performance,but IC-Kmeans significantly reduces complexity,especially at lower signal-to-noise ratios.Moreover,in indoor multipath channel models,NNIC-Kmeans outperforms IC-Kmeans and traditional K-means in terms of BER performance for optical orthogonal frequency division multiplexing(O-OFDM) systems. |
| Key words: visible light communication O-OFDM multipath channel channel equalization K-means algorithm BP neural network |
