| With the widespread deployment of wireless networks and the popularization of mobile smart terminals,multimedia applications such as web live broadcasts,sports event broadcasts and cloud games are booming,and the traffic of image/video-based multimedia services has also increased significantly.However,the highly dynamic wireless environment,heterogeneous terminal equipment,and massive data traffic have brought severe challenges to the design of wireless multimedia transmission mecha-nisms.The existing transmission mechanisms,including analog transmission and dig-ital transmission have achieved significant performance in their respective historical stages of development.However,there are still some issues in the actual deployment of these two transmission mechanisms.Analog transmission has shortcomings such as susceptibility to interference,chromaticity distortion,brightness crosstalk,and low definition,which affect user perception experience.Digital transmission was developed based on Shannon's source-channel separation theorem.In point-to-point communica-tion and the precise channel conditions are known,digital transmission can achieve opti-mal transmission performance.However,when the estimated channel condition and the actual channel condition are not match,"cliff effect" and "saturation effect" will occur in digital transmission,which will seriously affect system performance.Meanwhile,in the multicast and broadcast scenario,to ensure that all users can decode correctly,dig-ital transmission will adopt a conservative code rate,which leads to inefficient channel utilization and it is unfair for multiple users.Robust transmission is a novel transmission mechanism that solves the aforemen-tioned issues of analog transmission and digital transmission.Robust transmission re-moves the nonlinear operation in digital transmission.After decorrelating the original image and video,power scaling is directly performed on the decorrelation coefficient,and then the scaling coefficients are mapped to the physical layer symbols'quadrature and in-phase components for transmission.Since the entire system is linear,robust transmission can fundamentally remove the "cliff effect" and "saturation effect" in dig-ital transmission.Adopting appropriate resource allocation in robust transmission can effectively combat channel fading and noise.However,the current resource allocation mechanism in robust transmission is still in its infancy,and optimal performance cannot be achieved in some complex scenarios.In view of the importance of resource allocation in robust transmission and the deficiencies of existing research,this dissertation studies resource allocation mechanism in robust transmission,such as in bandwidth-constrained and heterogeneous resolution scenarios,as well as considering next-generation commu-nication technologies,such as large-scale antenna arrays millimeter wave to improve ro-bust video transmission efficiency and dynamic channel adaptability.This dissertation mainly includes the following contents:1)A compressive robust transmission mechanism is proposed for remote sensing image transmission in bandwidth-constrained scenarios.In the bandwidth-constrained scenario,the simple block discarding strategy in the original robust transmission is no longer suitable for remote sensing images.Different from natural images,remote sens-ing images contain more structural information.After decorrelation transformation,its energy is no longer concentrated in the low-frequency part.Instead,it is scattered throughout the whole frequency band.To address this issue,this dissertation intro-duces compressive sensing(CS)technology,and designs a robust transmission mech-anism for remote sensing images in bandwidth-constrained scenarios,to make full use of the sparseness of the medium and high frequency coefficients in the remote sensing image's transform domain.The basic strategy is to directly transmit these more im-portant coefficients using robust transmission,and these less important coefficients are first processed using CS and then transmitted utilizing robust transmission.Consider-ing the source distortion caused by the introduction of CS,we first analyse the system distortion,and then construct the system distortion minimization problem with respect to bandwidth and power allocation.For such mixed integer continuous programming problem,this dissertation proposes a two-stage resource allocation algorithm to solve the problem with low complexity.Simulation results show that CS can make full use of the sparsity of the middle and high frequency coefficients in the transform domain,and the proposed two-stage resource allocation algorithm can effectively improve system performance in bandwidth-constrained scenarios.2)A robust video transmission mechanism is proposed to accommodate users' het-erogeneous resolution requirements in multi-carrier broadcast scenarios.The problem of user resolution heterogeneity in broadcast scenarios has not been considered in robust transmission.To solve this problem,this dissertation first proposes a spatial decomposi-tion method to support differentiated resolution requirements of multiple users,as well as ensure the effective use of transmission bandwidth.Meanwhile,considering the di-versity of subcarriers' channel gain in OFDM system and the diversity of source chunks'importance,we formulate the problem of minimizing transmission distortion based on subcarrier matching and power allocation.This problem is a mixed integer nonlinear programming problem,which is difficult to obtain the optimal solution directly,thus we consider subproblem decomposition.For the power allocation subproblem,through the conversion of multi-user distortion coupling,we can derive the optimal closed-form expression of the power scaling factor.For the subcarrier matching subproblem,it is modeled as a bidding problem and we develop a bidding algorithm to get the ?-optimal solution with low time complexity.Then alternate optimization is utilized to solve the main problem and we give the algorithm convergence analysis.Meanwhile,an analog channel feedback mechanism is proposed to reduce the channel feedback overhead in broadcast scenarios.Simulation results show that the proposed scheme can support res-olution heterogeneity in the broadcast scenario,and the proposed resource allocation scheme can achieve approximate optimal performance with low computational com-plexity,as well as the alternate optimization algorithm can converge quickly.3)A robust transmission mechanism based on the large-scale antenna array mil-limeter wave(mmWave)is proposed for large volume video transmission.Large-scale antenna array based mm Wave is a potential technology to improve bandwidth effi-ciency and support large volume video transmission.By adopting the mm Wave lens antenna array,sparse mm Wave channel can be converted into mutually independent spatial paths,thus the processing complexity of mm Wave communication can be effec-tively reduced.Then utilizing the diversity of video content in application layer and the diversity of spatial path gain in physical layer,this dissertation designs a robust video cross-layer transmission mechanism.Since the number of spatial paths is usually less than the number of chunks,chunk scheduling is required under dynamic transmission environment,and path matching and power allocation should be considered to obtain optimal transmission performance.As the formulated distortion optimization problem is a mixed integer nonlinear programming problem,we consider subproblem decom-position to obtain the chunk scheduling problem,path matching and power allocation problem respectively.Then,based on the contribution of these chunks that have not been transmitted to packet distortion,we design a low-complexity chunk scheduling al-gorithm.Meanwhile,a greedy water-filling algorithm is designed based on the greedy idea to solve the path matching and power allocation problem.Simulation results show that the proposed scheme achieves better performance than the reference schemes in terms of objective peak signal-to-noise ratio and subjective visual quality.It also shows that the effectiveness of chunk scheduling and resource allocation,which can provide guiding suggestions for actual system deployment. |
PDF Full Text Request