Rate-Adaptive Schemes and Capacity Issues in Wireless Systems
MetadataShow full item record
The present dissertation consists of a collection of five papers and an introduction The papers are dedicated firstly to the analysis and design of efficient wireless communication systems operating over fading channels, and secondly to the information theoretic characterisation of fading channels under different sets of assumptions. The wireless systems that are considered operate over frequency-flat, block-fading channels under the assumption that perfect channel state information (CSI) is available both at the transmitter and at the receiver. One of the issues that is addressed is the energy-efficient design of constant power link adaptation schemes in which circuit energy consumption is included in the total energy budget, and in which the system's instantaneous bit-error rate must never exceed a predefined threshold. This is done assuming that a number of different transmission schemes are available for use, and that the objective is to maximise the system's average spectral efficiency (ASE). Another issue is how to optimally choose the rates of a given finite number of capacity-achieving codes in order to maximise the ASE of dual-branch multiple-input multiple-output (MIMO) systems (that is to say, MIMO systems with either two transmit or two receive antennæ). Yet another is the derivation of expressions for the signal-to-noise ratio (SNR) distributions of the independent subchannels which can be obtained by decoupling a dual-branch MIMO channel using linear precoding and decoding, assuming that the available power is distributed among these subchannels in accordance with the water-filling solution. It is also shown that the knowledge of the subchannel SNR distributions can be very helpful when optimising MIMO communication systems. The information theoretic characterisation of fading channels under different sets of assumptions is the other main concern of this thesis. The first channel that is considered from this perspective is the frequency-flat block-fading channel discussed above. The power adaptation strategy that maximises the average information rate that can be reliably transmitted over such a channel assuming the availability of perfect transmitter and receiver CSI, assuming that any transmitted codeword spans a single fading block, and assuming the input signal is subject to both average and peak power constraints is characterised by means of a theorem and numerically computed in different scenarios. The next channel that is considered is the memoryless noncoherent Rayleigh fading channel, in which the channel state is assumed to change on a symbol-by-symbol basis. Closed-form expressions for the mutual information between the output and the input of this channel when the input magnitude distribution is discrete and restricted to having two mass points are derived, and it is shown how these expressions can be used to obtain closed-form expressions for the capacity of this channel for SNR values of up to approximately 0 dB. The final channels to be examined are noncoherent Rayleigh-fading channels with memory, in which the channel state is once again assumed to change on a symbol-by-symbol basis, but where the fading process has memory which is modelled by an autoregressive process of arbitrary order. For such channels, it is shown that for any input magnitude distribution, it is optimum from a capacity perspective to choose the phase of the input independent and identically distributed, with a distribution which is uniform over the interval [0,2π).