Delay Analysis And System Design In Future Cellular Mobile Communication Systems

As a main part of the support system for mobile Internet and internet of things, the future cellular mobile communication systems should have the ability to meet the various performance requirements specified by diverse services. Delay performance guarantee will be one of the key capabilities in the future cellular mobile communication systems. However, the new technologies adopted by the future cellular mobile communication systems have complex effect on delay perfor-mance. For example, cognitive radio technology will make channel resource change frequently, and heterogeneous cellular network technology will result in highly random network topology. More-over, the delay performance usually contradicts with other system performances. For instance, improving delay performance requires more network construction and operation costs. In this dis-sertation, delay performance, tradeoffs between delay and other system performances, and system design for supporting delay-sensitive services are studied respectively for the future cellular mobile communication systems adopting the new technologies.For multi-channel cognitive radio systems, channel access technology for delay-sensitive ser-vices is studied, a channel access scheme is designed, and the performance of the scheme is anal-ysed. A two-phase channel access scheme is proposed, which consists of a distributed channel negotiation phase and a hopping-based channel access phase. An algorithm is designed for gen-erating the common hopping sequence used in the hopping-based channel access phase. For the distributed channel negotiation phase, the negotiation process is modelled as a two-dimensional Markov process, then transformed into a one-dimensional Markov process, and finally transformed into an absorbing Markov process. Based on properties of the absorbing Markov process, the av-erage maximum waiting time for channel negotiation is derived. For the hopping-based channel access phase, the effective capacity of the equivalent channel for each secondary user's service process is derived. Based on this, the maximum effective capacity and number of supportable sec-ondary users (SUs) per channel are defined, which can be used for call admission control design and system spectrum efficiency evaluation in cognitive radio systems.For heterogeneous cellular networks which are composed of traditional macro cells and dense small cells, the delay performance and the tradeoffs between delay and other system performance are investigated. By modelling the locations of the base stations and users in multi-tier cellular networks as independent Poisson point processes (PPPs), a framework for delay analysis in het-erogeneous cellular networks is established. The key system parameters and external environment factors are included, e.g., BS densities, user density, BS association parameter, loads in each tier, user mobility, and the temporal correlation of transmissions, et al. The local delay (the delay to transmit a packet with 100% reliability), user and area timely throughput (the mean number of suc-cessfully transmitted packets within the deadline) of heterogeneous cellular networks in static and high-mobility scenarios are derived using the framework, and then some meaningful insights are obtained. The results explicitly reveal the effects of the environment factors and key system pa-rameters on delay performance, and on the tradeoff between delay and reliability. An optimization problem is formulated to find the optimal tradeoff between local delay and network cost for dense homogeneous cellular networks, and then the optimal BS density is derived. The above theoretical work can provide guidelines for network design, especially the BS association parameter and BS density settings.For the future cellular mobile communication systems employing heterogeneous backhaul technologies, the delay performance and the cost of backhaul are studied. For the four most promis-ing backhaul technologies, namely, fiber, xDSL, sub-6 GHz and millimeter wave, their delays and costs are modeled, respectively. Using the delay models, the mean packet delays and delay-limited success probabilities of the four types of the backhauls are derived. Using the cost models, the average backhaul cost per small cell BS are studied, and the expression of the optimal gateway density is given which minimizes the average backhaul cost. As an application of the model, a BS association optimization problem is formulated for a two-tier cellular network with macro and small cells, which targets to minimize the mean network delay. The optimal bias factors in the BS association problem are derived. The above theoretical research and verification results also reveal the necessity of a joint design of radio access and backhaul networks for performance improvement.