Development Of Millimeter-wave Full-duplex Link-level Simulation Platform And Research On Digital Self-interference Cancellation Algorithms

At present,half-duplex mode is adopted in 4G communication,and the communication frequency band is concentrated on 6GHz and below.Due to the limited wireless bandwidth and the shortage of spectrum resource,how to improve the spectrum efficiency and information transmission rate to meet the large-scale data interaction with high-speed in the future 5th Generation has become an urgent problem to be solved.At the same time,half-duplex(CCFD)can improve the spectrum efficiency twice as much as that of co-frequency Co-time Full Duplex(CCFD)in theory,which is the future development vision of duplex technology.It can make full use of both frequency and time resource,and has the potential ability to improve the network spectrum efficiency.It is suitable for many communication scenarios such as spectrum shortage and fragmentation.Abundant free spectrum resource is available in millimeter-wave band,which can effectively meet the requirements for higher capacity and speed in the future 5G,and can support the user transmission rate above lOGbps.Therefore,it is of great significance to study millimeter-wave full-duplex for improving spectral efficiency and information transmission rate.The topic of this thesis comes from Beijing Natural Science Foundation:"Key Technologies Research in Next Generation Millimeter Wave Dense Cellular Networks"(Project No.4182955)and the Youth Project of China Institute of Information and Communication:"Research on the Performance of Related Algorithms on Physical Layer for 5G Millimeter Waves".Due to the high cost of millimeter-wave hardware devices,it is impossible to build a low-cost universal millimeter-wave hardware simulation platform,which limits the flexibility of research.Moreover,self-interference cancellation is the key to realize millimeter-wave full-duplex,and digital self-interference cancellation is particularly important in self-interference cancellation of millimeter-wave full-duplex,while the existing digital self-interference cancellation algorithms have limitations and poor performance.Therefore,in view of the communication requirements of millimeter-wave full-duplex,this thesis designs and implements a millimeter-wave full-duplex link-level simulation platform,and proposes a Long Short-Term Memory(LSTM)network for self-interference cancellation.The main contents of this thesis are as follows:1.Design and implementation of millimeter-wave full-duplex link-level simulation platformIn order to evaluate the performance of digital self-interference cancellation and system throughput rate in millimeter-wave full-duplex communication,the overall design of the platform contains process control module,transmitter module,channel module and receiver module.In the design of transmitter and receiver,millimeter-wave radio frequency device model is adopted,and millimeter-wave transmission characteristics are considered in the channel module.Considering the typical scene of millimeter-wave full-duplex communication,the millimeter-wave full-duplex communication scenarios including urban micro-cell,urban macro-cell and rural macro-cell are simulated in the platform,which is used to test,evaluate and verify the performance of millimeter-wave full-duplex communication.2.Design and implementation of the digital self-interference cancellation algorithm based on LSTMAiming at the enhanced multipath component in the digital self-interference signal in millimeter-wave full-duplex and the non-linear factor introduced by millimeter-wave devices,and considering the characteristics of time series of digital self-interference signal with multipath conditions,combined with the application scenario of LSTM neural network,a digital self-interference cancellation algorithm based on LSTM neural network is designed and implemented.Based on the platform,the performance of algorithm and system throughput rate with the proposed algorithm and the existing ones are compared.The proposed algorithm has good performance of digital self-interference cancellation and effectively improves the performance of millimeter-wave full-duplex communication.The simulation results show that compared with the traditional algorithms,the proposed algorithm improves the digital self-interference cancellation by 2.52 dB and the system throughput rate by 36.8%when the transmission power is 19 dBm.