Instantaneous And Time-variant Interference Modeling And Analysis Of Wireless Network

The development of wireless communications has greatly changed people’s life style. As the emergence of large scale wireless communication systems, such as ad hoc network, sensing network and heterogeneous network, the spatial structure of UEs and access points is usually far from being regular. Since the signal and the interference critically depend on the spatial distribution of transmitters, the network geometry and its structural fluctuations are critical parameters that influence the performance of wire-less networks. In order to analyze and design the wireless system, mathematical models should be proposed to accurately model the spatial distribution of the nodes. Traditional regular grid cannot be used to accurately model the randomness in the deployment of large scale wireless network; meanwhile, the close-formed results cannot be derived. Stochastic geometry, especially the point process theory, provides an efficient mathe-matic tools to model the spatial distribution of nodes, and close-formed results could be derived. Recently, stochastic geometry is shown to be powerful in the analysis of fundamental limit of wireless communication systems.The content of this thesis can be classified into two categories:modeling and per-formance analysis of instantaneous interference and time-varying interference. Spe-cially, four sub-problems are considered:instantaneous modeling and analysis of in-terference coordination, instantaneous modeling and analysis of resource management, time-variant modeling and analysis with backlogged traffic, time-variant modeling and analysis with busty traffic. Our main contributions are as follows:1) We explore the effect of two kinds of interference management mechanisms, i.e., dynamic TDD in TD-LTE and RTS/CTS mechanism in WLAN. An important mo-tivation to study the dynamic TDD is to model the practical scenario, and evaluate the difference between the practical scenario and the point process model. According to the theoretical analysis, we proposed three kinds of distributed interference coordination al-gorithms based on the OI information. The results reveal that BS-BS measurement is useful to improve the performance of the network. We propose a marked point process model to characterize the spatial distribution of transmitters and receivers, and the effect of RTS/CTS mechanism is similar to the thinning in Matern’s hard-core process. We propose a general approach to derive the mean interference at the typical receiver, and the analysis can be applied to more complicated protocols.2) We model the spatial distribution of users and access points in the heterogeneous network. we model the distribution of femtocells as Poisson point process or Neyman-Scott cluster process and derive the distributions of signal-to-interference-plus-noise ratios, and mean achievable rates, of both nonsubscribers and subscribers. The estab-lished expressions are amenable to numerical evaluation, and shed key insights into the performance tradeoff between subscribers and nonsubscribers. The analytical results are corroborated by numerical simulations.3) We focus on the question that whether increasing randomness in the MAC, such as frequency-hopping multiple access (FHMA) and ALOHA, helps to reduce the effect of interference correlation. We derive closed-form results for the mean and variance of the local delay for the two MAC protocols and evaluate the optimal parameters that minimize the mean local delay. Based on the optimal parameters, we propose the def-initions of two operation regimes:correlation-limited regime and bandwidth-limited regime. If no MAC dynamics is employed, the local delay has a heavy tail distribution which results in infinite mean local delay; meanwhile, simply employing FHMA and ALOHA will greatly decrease the mean local delay. By comparing the results of FHMA and ALOHA, we observed that while the mean local delays of the two protocols are the same for certain parameters, the variances are rather different. This work is an impor-tant reference for the study of interference correlation and for the design of practical system to optimize the delay.4) We investigated the stable packet arrival rate region and the delay of the dis-crete time slotted ALOHA network with the transmitters and receivers distributed as a Poisson bipolar process. The network is assumed to be static, and each transmitter in the network maintains a buffer of infinite capacity to store the incoming packets. We employed the tools from queueing theory as well as the point process theory and proposed several novel approaches to study the stability and the delay of this system by the simple concept of dominance. We obtained sufficient conditions and necessary conditions for stability as well as the upper bound and lower bound for the cumulative distributed function of delay in closed form. The numerical results showed that the gap between sufficient conditions and necessary conditions as well as the upper bound and lower bound are small and also revealed how the conditions and the bounds vary with different system parameters. To the best of our knowledge, this is the first work to study the stable packet arrival rate region and the delay of the static Poisson network. This work extends the interacting queueing problem to the overall network and proposed a novel problem, which is a good reference for the analysis combining queueing theory and information theory.