Reserch On Multi-service Mixed Transmission Optimization

Facing 5G evolution and 6G development,the services supported by mobile communication systems have diverse requirements in terms of capacity,throughput,delay,reliability,and connection density.The International Telecommunication Union divides 5G services into three major application scenarios:enhanced mobile broadband(eMBB),ultrareliable low-latency communication(uRLLC)and massive machine-type communication(mMTC).In B5G and even 6G mobile communication networks,the coexistence of multiple services with different needs will be very common,and there is an urgent need for mobile communication systems to provide differentiated on-demand and flexible services.It is difficult for a traditional single waveform to meet the needs of diverse services at the same time.Considering that differentiated services are provided for multiple business types on a unified physical layer,the physical layer must have a highly flexible framework.Waveform design based on mixed parameter sets can simultaneously carry network services of different business types by using waveforms with different parameters,thereby meeting the differentiated requirements of different services in various scenarios.However,due to the different intervals between the subcarriers between the waveforms of different parameter sets,the orthogonality between different waveforms will be destroyed,resulting in Inter-Numerology Interference(INI),which affects the performance of mixed services.service performance.Aiming at the above problems,this thesis conducts a multi-service-oriented hybrid transmission optimization research.The specific work is as follows:First of all,aiming at the problem of INI caused by out-of-band leakage(OOBE)in the mixed parameter set system,this thesis models and analyzes the causes of INI,and then analyzes the factors that affect its size and the factors that slow down INI through the expression of INI.The method provides a framework model for the following research.In order to solve the problem of INI in the mixed parameter set system,this thesis applies F-OFDM to the mixed parameter set scenario,and uses subband filtering technology to reduce OOBE.The simulation shows that subband filtering can effectively reduce OOBE and subband data bit error rate,At the same time,because different application scenarios have different requirements for transmission waveform configuration,and currently F-OFDM sub-band filters are generally seldom designed according to the specific application scenarios of each sub-band.Since different services in sub-bands are configured with waveforms of different parameters,the size of the INI between sub-bands will also be different,and the channel gain and channel noise of the mobile radio channel will have different influences on different parameter sets.This thesis proposes a heuristic algorithm-based mixed parameter set subband filter optimization strategy,which dynamically selects the filter parameters of each subband according to the configuration of subband waveform parameters and channel conditions to achieve the goal of maximizing the spectral efficiency of the system.Simulation results show that the proposed algorithm has better performance than existing methods.Then,aiming at the problem that bandwidth allocation causes the INI between B WPs to be aggravated by different parameter set powers between adjacent Bandwidth Parts(BWPs),thereby reducing system performance,this thesis proposes a transmission rate optimization scheme based on spectrum allocation.The purpose is to maximize the transmission rate of the total system users under the constraint of guaranteeing user service QoS through reasonable bandwidth allocation.Considering that the total bandwidth is divided into multiple sub-channels and that different services on sub-channels will interfere with each other,this thesis introduces a multi-branch dueling Q-Network(Branching Dueling Q-Network,BDQ)that can be used to deal with excessive action space.Solve optimization problems.Considering some data packets that do not meet the business QoS,this thesis adds an action masking module to the BDQ algorithm to improve the learning efficiency,Finally,the simulation results show that the proposed algorithm has better performance than the comparison algorithm in the case of different numbers of users.