Channel Estimation By Tensor-based Noise Compensation For Relay-Assisted Indoor VLC System

Light Fidelity(LiFi)technology,as a new communication technology that can supplement the shortcomings of traditional electromagnetic wave communication technology,has been proven to be effective in some specific communication environments.Compared with the traditional electromagnetic wave communication technology,visible light communication technology not only improves the transmission rate,but also improves the accuracy and confidentiality of information transmission,which makes it have a better advantage in the indoor transmission environment.However,since the channel state information is unknown in the general indoor information transmission environment,it needs to be obtained through channel estimation in order to get the best received signal.Different from the traditional electromagnetic wave communication system,the indoor visible light communication system needs to meet the non-negative and real value transmission data.Because the optical transmission signal is received by the photodiode,in the transmission data obtained by the receiver,in addition to the noise interference in the environment,it may also contain signals from other light sources,making the process of channel estimation more complicated.In order to solve these problems,this paper proposes a channel estimation algorithm based on the noise compensation tensor mode.The main research contents are as follows:Due to the most of the existing channel estimation algorithms use a pilot insertion method with high estimation efficiency.However,the current channel estimation algorithms based on the pilot method have not been able to design the transmission pilot based on the unique data structure of the sender of the indoor visible light communication system,which results in low estimation efficiency.The channel estimation method proposed in this paper,according to the principle of similar interference at the same frequency at the same time,a transmission pilot structure suitable for indoor visible light communication system is designed.Under the condition of satisfying the transmission conditions,the pilot structure can maximize the number of information transmission in a unit times improve the transmission efficiency,and get two types of pilot data used for channel estimation and noise estimation at the receiver.The method of multi-dimensional digital signal processing using tensor decomposition has been proved to be effective.According to the analysis of the data transmitted by the visible light communication system,it is proved that it is sending a multi-dimensional digital signal.Therefore,a corresponding suitable tensor decomposition structure is selected through a known structure of an indoor visible light communication system,and the noise estimation pilot and the channel estimation pilot are extracted at the corresponding positions on the receiving end to form a noise estimation tensor and a channel estimation tensor.The noise component in the channel estimation tensor is removed by using the noise estimation tensor,and the noise compensation estimation process of each parameter in the channel is realized by combining the tensor decomposition method.Based on the analysis of the actual indoor communication environment,the tensor model for various indoor visible light communication systems is established,such as one-way double-hop relay-forward indoor visible light communication system,relay-forward indoor visible light interfered by different light sources,and full duplex indoor visible light communication system.Then,we use the estimation method based on noise compensation to estimate the channel of several different indoor visible light communication systems.In order to improve the accuracy and efficiency of the estimation,the parameters in the channel are estimated jointly,and the most accurate channel parameter estimation value is obtained by iterative method In addition,the noise component is used as an estimation parameter,and the value of the noise estimation tensor is continuously updated during the estimation process,so that the whole estimation process can converge quickly,and the estimation accuracy is improved while the number of iterations is reduced.