Research On Non-orthogonal Air Interface For Wireless Communication System

Due to the rapid increase of wireless services,the future wireless communication systems should be able to support massive connectivity and achieve high spectral efficiency.In order to further improve the system capacity,the researchers have broken the orthogonality restriction in traditional communication systems and developed non-orthogonal air interfaces,including non-orthogonal waveforms and non-orthogonal multiple access schemes.In orthogonal multiple access(OMA)systems,several non-orthogonal waveforms have been proposed to suppress the out-of-band(OOB)leakage and improve the spectral efficiency,Filter bank multicarrier(FBMC)is one of the most widely investigated non-orthogonal waveforms.In order to suppress the OOB leakage,FBMC employs specifically designed pulse-shaping filters whose time-domain length is several times of the symbol period,which introduces intrinsic inter-symbol interference.As a result,the detection of FBMC is more complicated than that of the traditional orthogonal frequency division multiplexing(OFDM).On the other hand,FBMC has high peak-to-average power ratio(PAPR)since it is a multicarrier waveform.However,due to the intrinsic interference,the traditional PAPR reduction methods for individual OFDM symbols are no longer suitable for FBMC systems.Besides non-orthogonal waveforms,the researchers have also proposed non-orthogonal multiple access(NOMA)schemes,where multiple users can be supported by the same resource block at the same time,to improve the spectral efficiency and the number of connectivities of a communication system.Power-domain NOMA(PD-NOMA)is a typical NOMA scheme and has attracted much attention in recent years.In PD-NOMA systems,multiple users can share one resource block and be multiplexed in power domain.At the receiver,successive interference cancellation(SIC)is applied in detection and decoding.As a result,when the resource allocated to one user is modified,the performance of others users occupying the same resource block will also change,which makes the resource allocation of PD-NOMA very complicated.In this thesis,the basic concepts of FBMC and PD-NOMA are first introduced.Then,the detection and PAPR reduction problems of FBMC systems,as well as the resource allocation of PD-NOMA systems,are investigated,respectively.The main work of this thesis can be summa-rized as the following four aspects:Firstly,the detection of FBMC is investigated.Generally,the linear zero forcing(ZF)and minimum mean squared error(MMSE)equalizers are employed in the detection of FBMC signals due to their low complexity.However,they are not optimal and their performance are limited.In order to improve the detection performance,the optimal maximum likelihood(ML)and maximum a posterior(MAP)criteria should be considered.Nevertheless,the traditional optimal Viterbi and BCJR algorithms have extremely high complexity in FBMC systems due to the two-dimensional(2D)interference.Therefore,a novel low-complexity factor-graph-based MAP detector is devel-oped to make a good tradeoff between performance and complexity.Moreover,this detector can be easily adopted in a turbo structure to further improve the performance of coded systems because of its soft-input-soft-output feature.Secondly,the PAPR reduction of FBMC is investigated.Similar to the traditional OFDM,FBMC has a high PAPR since it is a multicarrier waveform.Due to the intrinsic interference,modifying one symbol block will also affect the PAPR of other symbol blocks in FBMC systems.Therefore,the traditional PAPR reduction methods for individual OFDM symbols can no longer be used in FBMC systems.In this thesis,the convex optimization based signal distortion approach is proposed to reduce the PAPR of FBMC signals.First,the optimization problem of the PAPR reduction is formulated,where the PAPR of a frame of FBMC signals is minimized subject to the constraints of symbol distortion and free subcarrier power.Then,a penalty concave-convex procedure(P-CCCP)based algorithm is proposed to solve the problem.Simulation results show that the proposed algorithm can effectively reduce the PAPR.Thirdly,the resource allocation of PD-NOMA is investigated.In PD-NOMA systems,the outage probability is an important performance metric to evaluate the system performance.More-over,the system performance can be improved by proper resource allocation scheme.In this thesis,the user scheduling and power allocation are optimized to minimize the transmit power of a down-link PD-NOMA system subject to the outage probability constraint.Besides the traditional outage probability,an alternative one that redefine the outage behavior is also considered.Moreover,im-perfect SIC is adopted in order to make the problem more realistic.To handle the complicated resource allocation problems,the user scheduling and power allocation are decoupled and two-phase low-complexity algorithms are proposed.Simulation results show that the performance of the proposed low-complexity algorithms is near-optimal and the algorithm based on the alternative outage probability outperforms that based on the traditional one when the imperfect SIC signifi-cantly affects the decoding.Finally,the application of PD-NOMA in mobile unmanned aerial vehicle(mobile-UAV)net-works is investigated.In PD-NOMA systems,the SIC order is determined by the channel gains of users.However,in mobile-UAV networks,the signal detection order may vary during the flight of the UAV,which should be specifically considered in the system design.In this thesis,PD-NOMA is applied to multiplex different users in a downlink mobile-UAV network.Moreover,the power allocation and UAV trajectory are jointly optimized to maximize the minimum achievable average rate among ground users.In order to address the complicated problem,a penalty dual-decomposition(PDD)based algorithm is proposed,which guarantees to converge to a stationary solution.Simulation results show that the proposed algorithm outperforms the benchmarks.