Dynamic time division duplex and time slot allocation strategy for multimedia traffic in wireless applications

by 1971- Jeong, Wun-Cheol

Abstract (Summary)
This thesis deals with dynamic time division duplex (D-TDD) operation and time slot allocation (TSA) strategy for dynamic traffic pattern in fixed wireless applications. The nature of emerging multimedia traffic is that it consists of several classes, which require different quality-of-service (QoS), and the bandwidth between uplink and downlink is quite asymmetric and dynamic. Frequency division duplex (FDD) operation and static time division duplex (S-TDD) operation may result in poor frequency utilization for this type of traffic, since the transmission bandwidth of these operations is fixed. Meanwhile, dynamic time division duplex (D-TDD) operation can support such asymmetric and dynamic traffic robustly, by adapting its transmission bandwidth according to traffic pattern. Thus, significant statistical multiplexing gain can be obtained in D-TDD operation. However, D-TDD is vulnerable in terms of strong interfering signals coming from base stations in co-channel cells, while the reference cell is busy with uplink receptions. Since the strong interference presence is due to the unique feature of D-TDD, the statistical multiplexing gain, which is also obtained via the unique feature, may not be obtained. In this thesis, we develop an analytic model of signal-to-interference ratio (SIR) distribution in D-TDD systems. Using the analytic model, we show that the probability density function (PDF) of SIR distribution in D-TDD system has two peaks, while that in conventional S-TDD systems has one peak. From the investigation of PDF, we show how the strong co-channel interference severely degrades the performance of SIR outage iv probability in D-TDD systems, when omni-directional antennas are employed at base station (BS) sites and subscriber (SC) sites. Also, we estimate the spatial resolution in terms of antenna beamwidth to obtain the target outage probability, when a spatial filter is employed at BSs. Numerical results show that 3.4-degree beamwidth is required to obtain an outage probability of 1% at the threshold SIR value of 17 dB. To suppress the effect of strong co-channel interference in D-TDD systems, we develop a time slot allocation strategy (TSA), exploiting spatially distributed SCs’ location over the cell coverage, combined with cost effective sector antenna layouts. We propose two TSA algorithms: the MaxMin {SIR} and the Max {SIR} algorithms. The MaxMin {SIR} is an exhaustive search algorithm, which searches the best pair set of time slots and SCs among all possible combination sets. The performance of this algorithm is used as an benchmark. Meanwhile, MaxMin {SIR} searches the pair in a specific order, which reduces the computational complexity, substantially. Our simulation results show that both algorithms perform well for a fairly large range of dynamic traffic pattern and the difference in performance between the two algorithms is not noticeable. Also, we compare the performance of TSA to that of adaptive array antennas. Our simulation shows that the performance of D-TDD system employing TSA, combined with 15 sector antennas, is comparable that employing adaptive array antennas with 26 sensing elements. Finally, we compare the spectral efficiency of TDD systems for various frameloading conditions. In the analysis, we consider two types of modulation schemes: fixed modulation and adaptive modulation. Fixed modulation is suitable for the delay-sensitive traffic, while adaptive modulation is proper for the delay-tolerant traffic, since the v transmission rate in delay-tolerant traffic can be reduced and the data bits can be buffered, until the channel quality is improved. In fixed modulation system, it is observed that the spectral efficiency of D-TDD systems can be improved significantly by employing TSA strategy. For instance, as much as 9 times higher spectral efficiency is obtained by employing TSA strategy, when the traffic is highly dynamic and the frame is fully loaded. Meanwhile, in adaptive modulation system, the spectral efficiency of D- TDD systems does not improve significantly by employing TSA strategy. This is due to the fact that the objective of the proposed system is to maximize the minimum value of SIR, not to improve the SIR values over all uplink time slots. Our numerical results show that the merit of statistical multiplexing can be obtained for a fairly large range of frame loading, by employing the proposed TSA strategy, compared to S-TDD systems. vi
Bibliographical Information:


School:Pennsylvania State University

School Location:USA - Pennsylvania

Source Type:Master's Thesis



Date of Publication:

© 2009 All Rights Reserved.