| The fifth-generation wireless communication system plays an important role in the current society,which greatly affects various aspects such as industrial production,life and entertainment.As one of the key technologies,millimeter-wave communication technology has many advantages,but the problems caused by the propagation characteristics of millimeter-wave cannot be ignored.In order to deal with the problem that millimeter wave communication is easily blocked by obstacles and has weak penetration,the current popular intelligent reflecting Surface(IRS)technology can be used to solve it.IRS is composed of a large number of low cost passive reflective elements,each of which can reflect signals independently by controlling its amplitude and phase,so it can provide a reflection link to communication when the line of sight(LOS)link is blocked by obstacles.Since the IRS is added to the millimeter wave communication system,the channel situation becomes complicated,and because of the passive characteristics of the IRS,it has no signal processing capability,and finally the channel estimation of the IRS assisted millimeter wave communication system becomes a challenge.In order to solve the channel estimation problem of the IRS assisted millimeter wave communication system,firstly,a regular alternating least squares based on tensor CP decomposition(CP-RALS)scheme for the system was proposed in this paper.Since the received signal model in this system is a three-dimensional tensor,it is necessary to use tensor CP decomposition to reduce its dimension,then use the regular alternating least squares method to estimate the channel,and design the regularization parameters through convex optimization to further improve the convergence speed and avoid divergence.Then,in order to estimate the channel between the base station and the IRS and the channel between the IRS and the users,the IRS structure is redesigned,that is,a hybrid IRS structure assisted millimeter wave communication system is constructed.Specifically,the hybrid IRS structure is to select some passive components on the IRS and equip each with a radio frequency(RF)chain.For this model,a channel estimation scheme based on an improved multiple signal classification algorithm and a complex parallel deep neural network is proposed.Among them,the improved multi-signal classification algorithm can estimate the departure angle and the arrival angle at the same time,while the complex parallel deep neural network considers the influence of the phase and sets a threshold value to further improve the estimation accuracy.Simulations show that this scheme can achieve better estimation accuracy.Finally,double-IRS assisted millimeter wave communication system was studied.Since mainly studies the channel estimation problem on the double reflection link,so it is proposed that the IRS on the base station side adopts a fully passive structure,and the IRS at the users side adopts a hybrid IRS structure.For this model,a two stage channel estimation scheme is proposed.In the first stage,the channel between the base station and the IRS at the base station and the channel between the IRS at the base station and the IRS at the users side are concatenated into one channel for estimation;In the second stage,the channel between the IRS on the users side and the users is estimated.According to the above solution,the complex channel estimation problem is decomposed into the problem of estimating the concatenated channel and the problem of estimating the channel between the IRS on the user side and the users.In summary,a variety of situations have been studied for the channel estimation problem of the IRS system.Firstly,the single-IRS assisted millimeter-wave communication channel estimation is studied.Then,the hybrid IRS structure-assisted millimeter-wave communication system is studied to improve the estimation accuracy.Finally,in order to cope with the complex communication environment double-IRS assisted millimeter wave communication system is studied. |
PDF Full Text Request