Application of auto-associative neural networks to transient fault detection in an IC engine

The tailpipe emissions from automotive engines have been subject to steadily reducing legislative limits. This reduction has been achieved through the addition of sub-systems to the basic four-stroke engine which thereby increases its complexity. To ensure the entire system functions correctly, each system and / or sub-systems needs to be continuously monitored for the presence of any faults or malfunctions. This is a requirement detailed within the On-Board Diagnostic (OBD) legislation. To date, a physical model approach has been adopted by me automotive industry for the monitoring requirement of OBD legislation. However, this approach has restrictions from the available knowledge base and computational load required. A neural network technique incorporating Multivariant Statistical Process Control (MSPC) has been proposed as an alternative method of building interrelationships between the measured variables and monitoring the correct operation of the engine. Building upon earlier work for steady state fault detection, this paper details the use of non-linear models based on an Auto-associate Neural Network (ANN) for fault detection under transient engine operation. The theory and use of the technique is shown in this paper with the application to the detection of air leaks within the inlet manifold system of a modern gasoline engine whilst operated on a pseudo-drive cycle. Copyright © 2007 by ASME.

Neural network based control for synchronous generators

In this paper, a Radial Basis Function neural network based AVR is proposed. A control strategy which generates local linear models from a global neural model on-line is used to derive controller feedback gains based on the Generalised Minimum Variance technique. Testing is carried out on a micromachine system which enables evaluation of practical implementation of the scheme. Constraints imposed by gathering training data, computational load, and memory requirements for the training algorithm are addressed.

The Synergistic Organization of Muscle Recruitment Constrains Visuomotor Adaptation

Reaching to visual targets engages the nervous system in a series of transformations between sensory information and motor commands. That which remains to be determined is the extent to which the processes that mediate sensorimotor adaptation to novel environments engage neural circuits that represent the required movement in joint-based or muscle-based coordinate systems. We sought to establish the contribution of these alternative representations to the process of visuomotor adaptation. To do so we applied a visuomotor rotation during a center-out isometric torque production task that involved flexion/extension and supination/pronation at the elbow-joint complex. In separate sessions, distinct half-quadrant rotations (i.e., 45°) were applied such that adaptation could be achieved either by only rescaling the individual joint torques (i.e., the visual target and torque target remained in the same quadrant) or by additionally requiring torque reversal at a contributing joint (i.e., the visual target and torque target were in different quadrants). Analysis of the time course of directional errors revealed that the degree of adaptation was lower (by ~20%) when reversals in the direction of joint torques were required. It has been established previously that in this task space, a transition between supination and pronation requires the engagement of a different set of muscle synergists, whereas in a transition between flexion and extension no such change is required. The additional observation that the initial level of adaptation was lower and the subsequent aftereffects were smaller, for trials that involved a pronation–supination transition than for those that involved a flexion–extension transition, supports the conclusion that the process of adaptation engaged, at least in part, neural circuits that represent the required motor output in a muscle-based coordinate system.

A fast method for fuzzy neural modeling and refinement

In the identification of complex dynamic systems using fuzzy neural networks, one of the main issues is the curse of dimensionality, which makes it difficult to retain a large number of system inputs or to consider a large number of fuzzy sets. Moreover, due to the correlations, not all possible network inputs or regression vectors in the network are necessary and adding them simply increases the model complexity and deteriorates the network generalisation performance. In this paper, the problem is solved by first proposing a fast algorithm for selection of network terms, and then introducing a refinement procedure to tackle the correlation issue. Simulation results show the efficacy of the method.

Adult neural stem cells expressing IL-10 confer potent immunomodulation and remyelination in experimental autoimmune encephalitis

Adult neural stem cells (aNSCs) derived from the subventricular zone of the brain show therapeutic effects in EAE, an animal model of the chronic inflammatory neurodegenerative disease MS; however, the beneficial effects are modest. One critical weakness of aNSC therapy may be an insufficient antiinflammatory effect. Here, we demonstrate that i.v. or i.c.v. injection of aNSCs engineered to secrete IL-10 (IL-10–aNSCs), a potent immunoregulatory cytokine, induced more profound functional and pathological recovery from ongoing EAE than that with control aNSCs. IL-10–aNSCs exhibited enhanced antiinflammatory effects in the periphery and inflammatory foci in the CNS compared with control aNSCs, more effectively reducing myelin damage, a hallmark of MS. When compared with mice treated with control aNSCs, those treated with IL-10–aNSCs demonstrated differentiation of transplanted cells into greater numbers of oligodendrocytes and neurons but fewer astrocytes, thus enhancing exogenous remyelination and neuron/axonal growth. Finally, IL-10–aNSCs converted a hostile environment to one supportive of neurons/oligodendrocytes, thereby promoting endogenous remyelination. Thus, aNSCs engineered to express IL-10 show enhanced ability to induce immune suppression, remyelination, and neuronal repair and may represent a novel approach that can substantially improve the efficacy of neural stem cell–based therapy in EAE/MS.

Persistent measles virus infection of mouse neural cells lacking known human entry receptors

Aims: Infection of the mouse central nervous system with wild type (WT) and vaccine strains of measles virus (MV) results in lack of clinical signs and limited antigen detection. It is considered that cell entry receptors for these viruses are not present on murine neural cells and infection is restricted at cell entry.

Methods: To examine this hypothesis, virus antigen and caspase 3 expression (for apoptosis) was compared in primary mixed, neural cell cultures infected in vitro or prepared from mice infected intracerebrally with WT, vaccine or rodent neuroadapted viruses. Viral RNA levels were examined in mouse brain by nested and real-time reverse transcriptase polymerase chain reaction.

Results: WT and vaccine strains were demonstrated for the first time to infect murine oligodendrocytes in addition to neurones despite a lack of the known MV cell receptors. Unexpectedly, the percentage of cells positive for viral antigen was higher for WT MV than neuroadapted virus in both in vitro and ex vivo cultures. In the latter the percentage of positive cells increased with time after mouse infection. Viral RNA (total and mRNA) was detected in brain for up to 20 days, while cultures were negative for caspase 3 in WT and vaccine virus infections.

Conclusions: WT and vaccine MV strains can use an endogenous cell entry receptor(s) or alternative virus uptake mechanism in murine neural cells. However, viral replication occurs at a low level and is associated with limited apoptosis. WT MV mouse infection may provide a model for the initial stages of persistent MV human central nervous system infections.