This paper presents a method of predicting the bit-error rate (BER) performance of the global system for mobile communications (GSM) system as a function of the wideband channel parameters obtainable from measured or assumed power delay profiles. Although delay spread is found to be the most useful parameter for the prediction of BER performance, it has been shown that in some conditions delay interval (15 dB) can provide a more accurate prediction mechanism. Thus, based on the delay spread and the delay interval (15 dB), empirical prediction formulas are derived.

Time dispersion measurements have been conducted in the urban city centre of Liverpool at 1800 MHz. The measurement setup is based on the swept time-delay cross-correlation technique and has a time-delay resolution of 33 ns. The use of automatic gain control (AGC) circuitry in the RF stage permitted the observation of the spatial variation of the multipath power delay profile and RMS delay spread. The experimental work investigated the dependence of RMS delay spread on base antenna height, gain, downtilt and distance in the LOS and non-LOS regions of the microcell. The results obtained indicate that (a) the RMS delay spread is less than or equal to 520ns/280ns in the LOS/non-LOS regions, (b) that the delay spread increases with antenna height and with the use of a directional antenna but decreases with downtilt. Furthermore, the RMS delay spread results are compared with theoretical results obtained using geometrical optics (GO) and the uniform theory of diffraction (UTD).>

An analysis of bit error rate (BER) in a binary digital FM system with postdetection diversity is presented. Expressions for the average BER due to additive white Gaussian noise (AWGN), random FM noise and delay-spread in the multipath channel are derived for reception using differential demodulation (DD), and frequency demodulation (FD) assuming independent fading signals. Calculated results for MSK show that the BER performance is strongly dependent on the RMS-delay/bit-duration ratio and that the delay-spectrum shape is of no importance when the receiver predetection filter product is optimized for the effect of AWGN. The effect of fading correlation on the diversity improvement is also analyzed for a two-branch case with multiplicative Rayleigh fading signals. Expressions for the average BER due to AWGN and random FM noise are derived. Calculated results are shown for the average BER due to random FM noise assuming a horizontally spaced antenna system at a mobile station. It is shown that the use of small antenna spacings leads to a diversity improvement greater than that obtainable for the case of independent AWGN.>

A method is proposed for predicting the degradation in performance of FSK data communication systems subjected to impulsive noise. It involves overlaying a representation of the incoming noise, expressed in terms of the noise amplitude distribution (NAD), on a family of isodegradation curves drawn with BER as the parameter. A direct prediction of the BER to be expected in the presence of the noise can be obtained by reading off the value of BER corresponding to the point of tangency between the two sets of curves (termed the tangential BER), and multiplying this by a factor F, characteristic of the radio location. The value of F can be found either experimentally or analytically. The usefulness of the method has been demonstrated, using vehicle-born equipment, by a series of experiments designed to be representative of typical mobile radio reception conditions. The method is very powerful experimental assessment technique. It is simple in concept, easy to apply, and has the advantage of being readily generalized for applications with different modulation techniques.>

Random FM noise using two-branch selection combining with correlated Rayleigh fading signals is analyzed for land mobile radio systems. General expressions are derived for the cumulative distribution function and mean square value of the random FM noise; they can be applied to any type of diversity such as space, polarization, etc. Calculated results show that, if two horizontally spaced antennas parallel with the direction of vehicle motion are used at a mobile station, random FM noise can be significantly reduced for small antenna spacings.>

A new method is proposed for assessing the performance of a radio receiver in the presence of impulsive noise. It involves selecting a criterion of performance, appropriate for the system under consideration (e.g., BER in data communication systems), and measuring isodegradation curves with respect to that criterion. A direct estimate of system performance can then be obtained using the noise amplitude distribution (NAD), one method being via the overlay technique. To demonstrate the validity of the method, three VHF/FM receivers have been tested in the laboratory. Results show that isodegradation curves in the new form give a better insight into performance degradation than those using the established criterion 20 dB quieting. It is suggested that receiver specifications should include a parameter related to performance in impulsive noise.