Designing fuzzy controllers from a variable structures standpoint

A procedure is presented for designing fuzzy controllers based upon variable structures techniques. Three such controllers are presented: the fuzzy equivalents of sliding-mode controllers, saturating controllers, and tanh controllers. By using an approach based upon variables structures (VSS) techniques, the stability of each of these controllers is assured. By using a sliding surface, the order of the rule base is reduced to size r×m, where r is the number of inputs and m is the number of fuzzification levels. This combination makes the proposed design procedure able to generate simple controllers with guaranteed stability properties. To illustrate the proposed design procedure, fuzzy controllers are designed for a ball-and-beam system. It is demonstrated that in spite of this system being a fourth-order unstable system, the proposed design procedure results in simple stable fuzzy controllers     

Comparison of analytic models for estimating the effect of clinical factors on the cost of coronary artery bypass graft surgery

The cost of treating disease depends on patient characteristics, but standard tools for analyzing the clinical predictors of cost have deficiencies. To explore whether survival analysis techniques might overcome some of these deficiencies in the analysis of cost data, we compared ordinary least square (OLS) linear regression (with and without transformation of the data) and binary logistic regression with two survival models: the Cox proportional hazards model and a parametric model assuming a Weibull distribution. Each model was applied to data from 155 patients undergoing coronary artery bypass grafting. We examined the effects of age, sex, ejection fraction, unstable angina, and number of diseased vessels on univariable and multivariable predictions of costs. The significant univariable predictors of cost were consistent in all models: ejection fraction was significant in all five models, and age and number of diseased vessels were each significant in all but the OLS model, while sex and angina type were significant in none of the models. The significant multivariable predictors of cost, however, differed according to model: ejection fraction was a significant multivariable predictor of cost in all five models, age was significant in three models, and number of diseased vessels was significant in one model. All five models were also used to predict the costs for an average patient undergoing surgery. The Cox model provided the most accurate predictions of mean cost, median cost, and the proportion of patients with high cost. This study shows: (1) lower ejection fraction and older age are independent clinical predictors of increased cost of CABG, and (2) the Cox proportional hazards model shows considerable promise for the analysis of the impact of clinical factors upon cost.     

Transforming the Engineering Curriculum: Lessons Learned from a Summer at Boeing

The Boeing Corporation conducts an A.D. Welliver summer fellowship program for engineering educators. This article describes the lessons learned by the 1998 summer Fellows. These include increasing emphasis on cost, communications and continuous learning, modifying faculty promotion guidelines to honor collaboration in teaching and research, and collaborating with industry on exit criteria. Eventually, industry has to become a partner in the educational process. The Fellows unanimously agreed that the Welliver Program was a valuable experience.     

Evaluation of technique to overcome small dataset problems during neural-network based contamination classification of packaged beef using integrated olfactory sensor system

Similar to most biological studies, beef contamination classification studies using artificial neural networks are restricted to small datasets. This study evaluates multivariate normal (MVN) technique of synthetic sample generation on small datasets associated with Salmonella contamination in beef. Six experiments were conducted to evaluate the performance of integrated sensor system towards identification of Salmonella contaminated beef packages. Pattern recognition involved using wavelet packet transform for feature extraction from sensor array responses and radial basis function network (RBFN) based classification of contaminated beef packages from uncontaminated packages. The MVN generated synthetic olfactory sensor signatures were used to train and test the RBFN classifiers. For the datasets analyzed in this study, genetic algorithm optimized RBF networks conferred average contamination test classification accuracies of 90.33% ± 7.68% (mean ± std. dev.) which were higher compared to the bootstrapped quadratic discriminant analysis based average accuracies. RBFN classifier based average overall classification accuracies of six synthetically generated datasets were in the range of 86.66%–98.89% with highest average overall classification accuracies of 98.89% ± 1.92%.     

Determinants of reoperation after 960 valve replacements with Carpentier-Edwards prostheses

During the period of 1977 to 1990, 960 Carpentier-Edwards standard prostheses (Baxter Healthcare Corp., Santa Ana, Calif.) were placed in 875 operations. Freedom from reoperation at 10 years was 57% +/- 4%, 76% +/- 3%, and 95% +/- 5% for mitral, aortic, and tricuspid valve replacement, respectively. Age was the only independent determinant of reoperation for both aortic and mitral valves. Likelihood of reoperation decreased with age, with freedom from reoperation after 10 years in patients aged less than 60 years versus 60 or more years being 65% +/- 5% versus 90% +/- 4% after aortic valve replacement and 48% +/- 5% versus 75% +/- 6% after mitral valve replacement. For mitral valve replacement, larger valve size made reoperation more likely, with freedom from reoperation at 10 years being 71% +/- 6% for sizes median less than 31 mm and 57% +/- 5% for sizes 31 mm or larger. For aortic valve replacement, prior median sternotomy reduced freedom from reoperation at 10 years from 80% +/- 3% to 25% +/- 5%. The low prevalence of reoperation affirms the suitability of the Carpentier-Edwards prosthesis for selected elderly patients and for tricuspid valve replacement. Because of their influence on the probability of reoperation, valve size and prior cardiac procedures also merit consideration in the choice of valvular prosthesis.     

Adaptation rates for repetitive control schemes

In this note the effect of the adaptation gains on the rate at which a repetitive control law forces an error to zero is presented. Once determined, a design procedure is suggested for how to choose the gains for the repetitive control law. © 1997 by John Wiley & Sons, Ltd.