Abstract: Signals representative of current and voltage in a circuit are processed and compared with stored temporal models models representative of arc faults and events not associated with arc faults. The models may be templates or stochastic models. Alternatively, the processed signals may be supplied to an artificial neural net programmed to recognize features of different arcs. An output may be provided to open a circuit breaker when an arc fault is detected.

Abstract: Multiple examples of a vocabulary of known words are recorded. The examples of each word are temporally aligned and supplied to a neural net which identifies unique features of each word which discriminate it from other words in the vocabulary. These unique features are integrated with parameters of a Hidden Semi-Markov model and stored. Speech signals from unknown words are then compared with the stored unique feature information after syntax restriction.

Abstract: Speech recognition is carried out by performing a first analysis of a speech signal using a Hidden Semi Markov Model and an asymmetric time warping algorithm. A second analysis is also performed using Multi-Layer Perceptron techniques in conjunction with a neural net. The first analysis is used by the second to identify word boundaries. Where the first analysis provides an indication of the word spoken above a certain level of confidence, an output representative of the word spoken may be generated solely in response to the first analysis, the second analysis being utilized when the level of confidence falls. The output controls a function of an aircraft and provides feedback to the speaker of the words spoken.