Diversity in cascaded Nakagami channels is studied. Average error probabilities and outage probabilities are calculated for dual selection combining algorithm in 2, 3, and 4-fold cascaded channels. Results demonstrate the existence of deterioration in the performance of wireless systems as the number of cascades increases from 1 to 4, and, the improvement gained through the implementation of selection combining diversity.

Short-term fading in wireless systems can be described using a cascaded approach permitting the modelling of severely faded channels. A maximal ratio combining (MRC) diversity is implemented to study the performance improvement in such channels. While the probability density function (pdf) of the signal-to-noise ratio of the independent and identically distributed N*Nakagami cascade channels is available in terms of Meijer's G functions, the pdf of the MRC output is not available. Because of this, the moment generating function of the MRC output is used to estimate the average bit error rates. The characteristic function is used to obtain the cumulative distribution (CDF) of the MRC output by applying the Gil-Pelaez approximation. The outage probabilities were then estimated using the CDF. Results showed that the wireless system performance improved with diversity even though the level of improvement declined as the number of cascaded components N increased. The analysis also suggests that additional mitigation such as combining signals from multiple base stations (macrodiversity) might be necessary to significantly improve the performances of cascaded channels.

Statistical analysis of Ultra-Wideband (UWB) signals is undertaken to examine the existence of short term fading and shadowing. Based on the data collected from multiple locations and 4 different sites using a single transmitter and a 96-element receiving antenna in the 2-8 GHz range, hypothesis testing was performed to explore the statistical fit to composite probability density functions such the double Rayleigh, double Nakagami and the K distribution as well as the Nakagami and lognormal. The tests clearly showed that excellent match existed with the K distribution, with the Nakagami distribution being the next best. These results point to the existence of shadowing besides the short fading reported by other researchers. They can also assist in the development of fading and shadowing mitigation techniques.

Speckle in ultrasonic image systems adversely impacts the contrast and resolution in the image. This poses serious problems in the interpretation of B mode images of internal organs such as breast, liver, kidney and so on. In the absence of sufficient contrast, classifying the regions of interest into benign and malignant masses becomes error prone. Since some of the masses are uniquely identified in terms of the boundaries, poor contrast and resolution will result in difficulties with their identification. A new class of spatial filters based on cylindrical Bessel functions of the first kind is proposed for speckle reduction. These filters with complex impulse responses were explored for enhancing the contrast of speckled images. Hypothesising that the phase of the filtered image carries boundary information, the phase characteristics of four speckled images are also studied for detecting boundaries. Results indicate that these filters do improve the contrast and enhance the boundaries. It is shown that the phase map clearly indicates the existence of boundaries. A simple thresholding applied to the phase highlights the boundaries. The results show the strength of the Bessel spatial filters in improving contrast and highlighting boundaries without resorting to any additional edge-detection algorithms.

Breast cancer diagnosis through ultrasound tissue characterization was studied using receiver operating characteristic (ROC) analysis of combinations of acoustic features, patient age, and radiological findings. A feature fusion method was devised that operates even if only partial diagnostic data are available. The ROC methodology uses ordinal dominance theory and bootstrap resampling to evaluate A(z) and confidence intervals in simple as well as paired data analyses. The combined diagnostic feature had an A(z) of 0.96 with a confidence interval of at a significance level of 0.05. The combined features show statistically significant improvement over prebiopsy radiological findings. These results indicate that ultrasound tissue characterization, in combination with patient record and clinical findings, may greatly reduce the need to perform biopsies of benign breast lesions.

Fiber feeders have been proposed for simplifying the base station in microcells, and thus make feasible the use of the microcellular architecture to overcome the barrier posed by the finite available spectra. We provide an analytical model to evaluate the exact error performance of the system, which accounts for the statistical effects in the mobile channel and the nonlinearity of the laser diode. The nonlinearity of the optical source coupled with the fading nature of the wireless channel mandates a solution based on simulation. However, the computer run time for the simulation is prohibitive, virtually rendering the model useless. We propose a computer-amenable "simplified model" which reduces the computation time immensely and also provides reasonably tight bounds. The BER performance for various channel conditions is evaluated using both models, and comparative studies are made on fiber-optic and non-fiber-optic systems. The tradeoff between the system performance and system capacity is discussed. Results obtained justify the use of fiber-optic feeders for future microcellular systems from a BER-performance point of view.

A model for the scattering of ultrasound from breast tissue is proposed. The model is based on the use of non-Rayleigh statistics, specifically the K distribution to describe the backscattered echo from the tissue. A multiparameter test based on this model has been designed to characterize the tissue. The data from the B-scan images of the breasts of 6 different patients were analyzed using this model. The results indicate that the non-Rayleigh statistics seem to be useful in characterizing and identifying malignant, benign, and normal tissue regions.