Rake receiver performance in direct sequence spread spectrum systems operating at high data rates

Work in this thesis includes a more comprehensive analysis than that available in the literature, of the problem of self-interference in direct sequence spread spectrum systems operating at high data rates. Performance estimates in terms of bit error rate as a function of signal-to-noise ratio at the receiver are derived for different Rake receiver configurations. Semi-analytical techniques previously reported in the literature by Lehnert and Pursley to analyse multiple-access interference in non-fading environments are extended to apply to the problem of estimating self-interference in fading environments. These techniques avoid the Gaussian approximation for the interference, which can lead to inaccurate performance estimates. Additionally, from channel impulse response estimates measured in urban microcellular environments it has been observed that time-delayed replicas of the transmitted signal have unequal average powers in general and the fading is not always uncorrelated. The semi-analytical approach is therefore also used to analyse performance for arbitrary fading conditions and compare estimated performance with results obtained when standard channel models that assume equal average powers and independent fading are used. Deficiencies in the results from application of standard models are reported.; Knowledge of performance deficiencies and reasons for them can be used for better allocation of resources such as transmit power for different operating modes that use, for example different spreading factors, and for different radio channel conditions, such as different numbers of multipath groups and their associated unequal relative average powers.; To enable coherent detection in many current systems using Rake receivers, a pilot sequence is code-multiplexed with the data for the purpose of estimating the complex amplitudes of received multipath groups. As a result of both self-interference and thermal noise, errors can occur in the estimates of these amplitudes. Again, the semi-analytical technique can be used to permit consideration of Rake estimation errors in the analysis of performance for different channel conditions and receiver configurations. An increase in power of the data signal can cause an increase in interference to the pilot channel thereby increasing the magnitude of errors on channel estimates, while an increase in the power of the pilot signal will reduce the channel estimation error while enhancing the interference to the data. The semi-analytical approach is used in the thesis to determine a suitable power level difference between the data and pilot signals. In cases where channel estimation errors are significant, analysis results indicate a degradation in performance with the inclusion of Rake fingers for multipath groups having powers below a certain threshold relative to the power of the strongest multipath group, concomitant with poorer channel estimates. Hence the requirement for a finger activation threshold. The resulting reduction in the number of active Rake fingers has the additional advantage of decreasing power consumption in the mobile terminal. (Abstract shortened by UMI.)...