Research On Tensor-based Channel Estimation And Hybrid Precoding For Millimeter-wave Massive MIMO Internet Of Vehicles

The millimeter-wave massive MIMO communication system uses beamforming to form a beam with high path gain in a specific direction to reduce the influence of multipath fading on the signal.In the high-speed environment,the channels of the vehicle users are more complicated due to the existence of Doppler frequency shift.Especially in the millimeter-wave massive MIMO car networking system,the channel fading is more obvious.Tensors can achieve lower computational complexity when dealing with high-dimensional data due to their ability to reduce the complexity of multidimensional arrays.In this paper,In this paper,aiming at the problem of improving the system stability in the high-speed environment of the millimeter-wave massive MIMO vehicle networking system,the tensors is used to study the vehicle networking channel estimation and hybrid precoding.First,in order to solve the problem of the rapid fading of the V2 X channel in highspeed environment,the channel estimation of V2 X communication cannot be accurately obtained,which leads to the reduction of system data transmission efficiency.A Bayesian tensor-based channel estimation scheme for V2 X millimeter-wave massive MIMO system is proposed.Firstly,a time-varying channel model of mm Wave Internet of Vehicles is established,and a super-resolution grid is designed to accurately estimate the signal arrival and departure angles.The millimeter-wave channel gain matrix and Doppler frequency shift can be accurately obtained by decomposing the low-rank tensor of the signal received by the vehicle user.In order to solve the non-real-time problem of the angle of arrival caused by the high-speed mobility of users,accurate real-time channel angle domain information is obtained by using the integrated sensing and communications of the angle of arrival.The simulation results show that compared with other algorithms,the proposed algorithm still has higher accuracy in high-speed environment.Second,due to the rapid fading of the V2 X millimeter-wave massive MIMO channel,accurate beamforming cannot be obtained,which reduces the system spectral efficiency.In this paper,a multi-vehicle massive MIMO hybrid precoding scheme is designed to solve the problem of frequent beam switching caused by vehicles.In order to maximize the user's SNR,a multi-vehicle joint optimization simulation beam alignment algorithm is designed to train the optimal simulation transmit matrix and combination matrix.In order to reduce the beamforming interference between users,the time dimension is extended to construct the equivalent baseband tensor channel of the user,and the digital precoding matrix is obtained by optimizing the signal-toleakage-to-noise ratio of the user.The digital combination matrix is obtained by minimizing the mean square error of the received signal.Simulation results show that the proposed algorithm outperforms other schemes in terms of system spectral efficiency.Finally,the above algorithm is verified experimentally by constructing a real road scene and a millimeter-wave massive MIMO vehicle networking communication system.For two common vehicle networking communication scenarios,the practicability of the above channel estimation and hybrid precoding scheme is carried out.The experimental results show that the proposed channel estimation and hybrid precoding scheme are robust to high-speed environments and perform better in terms of bit error rate and throughput.