RESUMO
Fault diagnosis plays an important role in the operation of modern robotic systems. A number of researchers have proposed fault diagnosis architectures for robotic manipulators using the model-based analytical redundancy approach. One of the key issues in the design of such fault diagnosis schemes is the effect of modeling uncertainties on their performance. This paper investigates the problem of fault diagnosis in rigid-link robotic manipulators with modeling uncertainties. A learning architecture with sigmoidal neural networks is used to monitor the robotic system for any off-nominal behavior due to faults. The robustness and stability properties of the fault diagnosis scheme are rigorously established. Simulation examples are presented to illustrate the ability of the neural-network-based robust fault diagnosis scheme to detect and accommodate faults in a two-link robotic manipulator.
RESUMO
This study reports the predictive value, in septic patients, of septic shock at presentation (SS factor) alone and in combination with multiple markers, using survival of the sepsis episode as the outcome measure. The SS factor correctly predicted the outcome in 53/68 (78%) of patients in this study. The Acute Physiology and Chronic Health Evaluation II Score (APACHE II or APII) and interleukin-6 (IL-6) and IL-6 soluble receptor (IL-6sR) concentrations were evaluated in combination with the SS factor in the same 68 patient population which was randomly divided into design (# = 50) and test groups (# = 18). Two iterations of an algorithm were evaluated using randomized patient groups corresponding to those producing the best (Group A) and worst (Group B) performance using a neural network. The four-input algorithm (APII, IL-6, IL-6sR, SS factor) correctly classified 16/18 (89%, Group A) and 14/18 (78%, Group B) of patients in the test subset. The corresponding four-input neural network model (10 iterations) correctly classified 61 to 89% of the 18 patients in the 10 test subsets.
