Research On Terahertz Joint Communication And Positioning Technology Based On Massive MIMO

Terahertz(THz)Joint Communication and Positioning System(JCPS),which depends on the THz’s high bandwidth and resolution of delay and angle and combines the Massive Multi-Input Multi-Output(Massive MIMO)antenna array technology,has been the new hotpot in 6G research.However,the THz user’s receival signal to noise ratio will fluctuate wildly due to the irregular movement of users and the lack of the power allocation in the practical application of JCPS,which will seriously affect the overall performance of the system.In order to stabilize the received signal,the thesis studies the beam tracking and robust power allocation algorithms in THz JCPS.The main research contents and innovations of the thesis are as follows.Firstly,a beam tracking algorithm based on Proximal Policy Optimization(PPO)is proposed aiming at the problem of channel fluctuations caused by user’s movements in THz JCPS.Firstly,the system model of THz JCPS is built and the learning environment of Deep Reinforcement Learning(DRL)is constructed.On this basis,the beam tracking network for mobile users is trained and the random path simulation test for the training results is carried out.Simulation results verify that the proposed dynamic beam tracking algorithm can effectively track the random mobile users and stabilize the fluctuating signal to noise ratio to about 12 d B under noise-174 d Bm/Hz,and has more 1.2 bps/Hz rate performance than the traditional grain-size search beam tracking algorithm under 4 times beam updating.Secondly,the internal relationship between communication and positioning is analyzed with Cramer-Rao Lower Bound(CRB)to solve the power allocation problem in THz JCPS.Then two robust power allocation algorithms are proposed.Firstly,the Lagrange penalty factor is used to obtain the efficient solution of robust power allocation through DRL.Secondly,Bornstein inequality is used to transform probabilistic constraints into deterministic constraints related to distribution parameters.Then the problem can be solved based on convex optimization.Simulation results verify that the proposed robust power allocation algorithm can achieve better positioning performance on the premise of guaranteeing the communication rate 160 Mbps of the system under 200 m distance.The lower bound of positioning error is 0.5m with proposed algorithm,which is 0.7 m lower than equal power allocation scheme.Finally,the thesis establishes a 3-dimensional THz JCPS system model for the problem of inapplicability of beam tracking algorithm in 3-dimensional senses.Combining with robust power allocation algorithm,the dynamic beam tracking algorithm of 3-dimensional scene is proposed and simulated.Simulation results verify that the proposed algorithm can effectively stabilize the receival signal to noise ratio of mobile users in 3-dimensional scenes,guarantee the service quality of mobile users,and improve the utilization rate of system resources and system performance.This thesis contains 31 figures,7 tables and 89 references in total.