Research On Multimedia Transmission Technologies In Wireless Networks Based On Cross-Layer Optimization

Along with the rapid development of wireless communication technologies and the proliferation of mobile smart devices, the traffic types in wireless networks are becoming more and more rich. The traditional traffic types like voice and messaging will be gradu-ally dominated by the wireless multimedia service such as the mobile video. The mobile video traffic tends to be exponential growth, which leads to the rapid increasing demands for the bandwidth and energy in wireless networks. There still exists many challenges for the wireless communication technologies to satisfy the explosive growth demands of multimedia traffic with the limited spectrum and power resources. The multimedia com-munications involve in the multiple layers of the communication protocol architecture such as the application layer, the network layer as well as the physical layer. The meth-ods based on the traditional layered protocol architecture cannot meet the demands of the multimedia communications due to the lack of an effective interaction among multiple layers. Cross-layer design (CLD) can achieve the optimal performance of the commu-nication system through jointly optimizing the system parameters from multiple layers. Therefore, to address the above challenges, CLD is a potential solution to improve the spectrum efficiency and energy efficiency of the entire wireless communication system.The objective of this dissertation is to design some effective multimedia transmission schemes for satisfying the demands for the quality-of-service (QoS) and the quality-of-experience (QoE) in the future wireless communication networks. To this end, from the idea of improving the utilization efficiency of radio resources, this dissertation focuses on investigating the resource optimization methods of wireless multimedia transmission, the interference-aware cross-layer design for distributed multimedia transmission as well as the optimal media access scheme in interference-limited wireless networks. Taking the video transmission for an example, this dissertation focuses on studying the video transmission schemes, theoretical analysis and mathematical modeling, algorithm design as well as the performance evaluation through utilizing the information theory, queue-ing theory, convex optimization theory, game theory and the stochastic geometry theory. Through investigation, this dissertation proposes wireless resource optimization methods, wireless video transmission schemes as well as an optimal media access strategy. The main contributions and novelties of the dissertation are summarized as follows.1) For the multi-users wireless video transmission networks, this dissertation inves- tigates the key problem that how to maximize the video transmission quality under the QoS-constraints. A CLD-based video transmission framework is developed by jointly considering the transmit power at the physical layer, queueing delay at the MAC layer as well as the video coding rate at the application layer. Then, a CLD-based radio re-source optimization method is proposed through jointly optimizing the transmit power and video coding rate. This method effectively solves the mismatching problem between the video coding rate and the channel transmission rate. In addition, this dissertation also analyzes the impacts of video coding rate, queueing delay as well as the transmit power on the video distortion model. The problem of maximizing the video transmission qual-ity is formulated into the optimization problem of minimizing the sum distortion of all users under the delay and power constraints. By utilizing the convex optimization theo-ry and the reformulation-linearization relaxation technology, the non-convex problem is converted into a convex one, which can be solved through utilizing the standard convex optimization technology. The performance of the proposed cross-layer assisted algorith-m is evaluated through simulations. For performance comparison, this dissertation also presents the simulation results of the artificial bee colony optimization algorithm and the fixed power optimization algorithm. Simulation results show that the proposed cross-layer assisted algorithm improves the video transmission quality considerably through the com-parison with the other two algorithms. These contributions correspond to the Chapter 2 and the paper 3 and 5 listed at the end of the dissertation.2) For the wireless ad-hoc networks, this dissertation proposes an interference-aware cross-layer design for the distributed video transmission. The central idea of this scheme is given as follows. Each session (a source-destination pair) first chooses one best frequen-cy channel and then decides whether to transmit the arrival packet or not by comparing its channel gain with a given threshold. If the channel gain is above the threshold, the packet will be transmitted, otherwise enqueued in the buffer. The main novelties are summarized in the following aspects:① A threshold-based transmission strategy is proposed for video transmission over a wireless ad hoc network to address the problem that how to choose channel and whether to transmit packet or not in a given slot.② Based on the proposed transmission strategy, this dissertation develops a novel in-terference approximation method to model the stochastic interference by consider-ing the dynamic network state. The accuracy of the interference model is validated through Monte Carlo simulation.③ Based on the queueing theory, this dissertation utilizes a M/M/1 model to model the queueing delay for the distributed network scenarios. Besides, a closed-form expression of the overall packet loss probability is also derived in this dissertation.④ This dissertation establishes an interference-aware cross-layer scheme for video transmission, based on which the problem of maximizing average video quality is formulated into a cross-layer optimization problem by jointly considering the video coding rate, the queueing delay as well as the transmission threshold. Considering the facts that the network scenario is distributed and the optimization problem is non-linear and non-convex, a distributed algorithm is proposed to solve the formulated optimization problem based on the convex optimization and game theory.⑤ Both the theoretical analysis and simulation results show that the proposed dis-tributed algorithm can improve the average video transmission quality and also con-verges quickly. The proposed algorithm can reduce the computation complexity and the signaling overhead. Therefore, this scheme provides an important theoretical re-search value and a practical application value in the future networks with low latency and high reliability requirements.These contributions correspond to the Chapter 2 and the paper 1 and 4 listed at the end of the dissertation.3) For the interference-limited wireless networks, this dissertation investigates the prob-lem of the QoS-constrained optimal media access probability (MAP) and proposes an op-timal media access strategy with the QoS guarantee to maximize the network throughput. The main novelties are summarized in the following aspects.① Based on the stochastic SINR physical model, this dissertation designs a random access protocol with the delay QoS guarantee. Then, the stochastic property of the aggregated interference is analyzed based on the stochastic geometry theory. Since the aggregated interference is of heavy-tailed property, its distribution is modeled as log-normal distribution. The accuracy of the log-normal model is verified through Monto-Carlo simulations.② Based on the proposed random access protocol, the total delay for transmitting one packet for a user is analyzed and derived based on the M/G/1 queueing model for the interference-limited networks. The problem of maximizing the network throughput is formulated into an optimization problem by taking into accounts the traffic char-acteristics, network topology, queueing delay, physical channel condition as well as the stochastic interference.③ Considering that the application-specific traffic conditions of a network, two cases with homogeneous user traffic and heterogeneous user traffic are discussed to solve the formulated optimization problem respectively. For the homogeneous user traffic, a closed-form expression of the optimal MAP is derived and a global optimization algorithm is proposed with low complexity.④ For the heterogeneous user traffic, considering the fact that it is difficult to solve the formulated non-linear and non-convex problem, this dissertation proposes a global optimal algorithm based on the branch and bound framework and convex relaxation technology. The proposed global optimal algorithm not only improves the network throughput performance, but also converges to the optimal solution with arbitrarily high accuracy after a limited number of iterations. Furthermore, Simulations results show that the proposed global optimal algorithm can improve the network through-put significantly. The algorithm offers a theoretical performance benchmark for the future protocol and algorithm design in the wireless interference-limited networks.These contributions correspond to the Chapter 4 and the paper 2 listed at the end of the dissertation.