Research On Delay Analysis And Resource Allocation For Large-scale Wireless Integrated Networks

With the rapid application of advanced technologies such as Internet of things(IoT),mobile edge computing(MEC)and machine learning in 5G/B5G wireless scenarios becoming more and more mature,the total amount of mobile service data around the world has increased dramatically.The full coverage of large-scale integrated networks and ultrareliable low-latency wireless guarantees have become popular research areas to improve the quality of service(QoS)and quality of experience(Qo E).Large-scale dense deployment of integrated networks,dense clustering in hotspot areas,and efficient multi-access techniques can effectively improve network capacity and instant transmission rate.Therefore,these technologies have become the core techniques of the next generation wireless networks.However,future large-scale integrated networks should be heterogeneous networks composed of feature-rich mobile terminals,of which the network and data statistics will be greatly cpmlicated,and the correspoding network-layer performance will be hard to be described.Meanwhile,the network-layer performance of the large-scale integrated network is an average metric that based on the performance of all links in the network.Thus,it seems difficult to describe the link-layer end-to-end(E2E)queuing delay and backlog.Hence,how to describe the network characteristics under various heterogeneous wireless networks? How to evaluate the link-layer queue status for large-scale integrated networks? The above questions need to be explored and answered thoroughly.According to the above discussion,the network-layer E2E statistical delay performance are investigated,based on the spatial distribution of the network devices and the statistical characteristics of the interference channels.Meanwhile,we study the performance difference under various multiple access techniques to explore the the performance enhancement under these techniques.Main contributions of this work are listed as:1.We have proposed an analytical method for network modeling and network-layer E2E delay violation upper bound analysis of large-scale communication networks,under different access strategies.Under certain network and data models,we have quantitatively evaluated the delay metric for fixed transmission pair and transmission under nearest association strategy.With the aid of stochastic geometry,the interference under spectrum reuse has been well modeled and quantificated for communication networks.Based on the interference capturing,the statistical characteristics of the instantaneous signal-to-interference-plus-noise ratio(SINR)can be well evaluated,with which we can further obtain the service capability and derive the network-layer delay violation upper bound via stochastic network calculus(SNC)in the exponential domain.The analytical framework reveals how the network parameters impact the E2E delay.Specifically,the derived mathematical expression shows the impact of important paramters on the network-layer delay violation including network parameters(e.g.,base station density and transmitting power,etc.),wireless channel parameters(e.g.,largescale fading and small-scale fading,etc.),data flow parameters(e.g.,arrival rate)and different access strategies.The results show that the delay violation bound of fixed transmission pair depends on the concrete transmitting distance,whilst the nearest association strategy can allocate network resources more conveniently and reduce the delay level efficiently.The analysis verifies the effectiveness of the proposed method for the E2E latency analysis of large-scale heterogeneous networks.2.This article has proposed efficient interference avoidance,delay analysis and resource allocation methods for an integrated computing and communication mobile edge computing(MEC)network.These methods have achieved joint gains of unified resource management and local-cloud fair scheduling.In such integrated network,the service capabilities of wireless communication and data computing are analyzed asynchronously.The delay of the two-hop first-offloading-then-computing queue of the integrated network are captured quantitatively under various multiple access techniques.Under certain network deployment,MEC technology can improve the resource utilization at the edge,and reduce the server load and data transmission latency efficiently.Meanwhile,we adopt various multiple access methods to further manage the inter-cluster and intra-cluster interference under spectrum reuse.Based on the concatenation property of the SNC approach,we capture the service capability of the equivalant queue model of the two-stage first-offloading-then-computing queue and evaluate the queuing delay for the MEC network.The research results show that intra-cluster orthogonal frequency division multiple access(OFDMA)strategy and OFDMnon-orthogonal multiple access(NOMA)strategy shows different delay characteristics for the network under different network parameters.This article has optimized the data offloading process and proposes the optimal resource allocation method to cater for a target delay requirement when considering the fairness between local and MEC processing resources.3.This article has proposed an analytical framework to analyze the transient delay under emergency that results in antenna misalignment for an integrated sensing and communication unmanned aerial vehicles(UAV)network.Under certain network deployment,we have quantitatively revealed how the steady-state queue backlog and UAV antenna control module influence the network performance.This work has captured the wireless fading channel characteristics of high-altitude millimeter wave(mm Wave)transmissions.The service capabilities of sensory data relaying and data uploading from sensing UAVs and fusion center UAVs have been derived based on the channel dynamics.Furthermore,the network-layer steady-state delay performance are captured for the sensory data transmission which can be denoted as a two-stage first-relay-then-uploading queue.Meanwhile,under the circumstance of relaying antenna misalignment,we evaluate the service capability of antenna side lobe transmission during several time slots.With the service capability,the delay impulse can be derived for sensing UAVs equipped with various control module under different antenna realignment standards.Particularly,we show that with further network densification,we can still achieve a huge increase in network capacity while maintaining the QoS requirement under advanced communication standards.The results are useful for practical network design and deployment.