Tone recognition in post-lingual deafened Nucleus cochlear recipients;

Curve fitting of pitch contour used for tone recognition of isolated mandarin syllables;

Approaches to pitch processing in tone recognition of Mandarin;

Automatic digital modulation classification algorithm based on novel combined feature vector

A new method of modulation classification for digital communication signals,using normalized fourth-order cumulants as characteristics and a covering algorithm for constructing neural networks as a classifier,is proposed.

Noises and fading channel are very important factors affecting the success rate of modulation classification in practical situation.

Study on efficient adaptive clustering algorithm based modulation identification;

Method of modulation identification for digital communication signals;

A blind equalization based algorithm of modulation identification in multipath channel enviroments;

Algorithm for modulation recognition in a time-variant multipath channel environment;

The application of support vector machine to digital signal modulation recognition system;

Hierarchical digital modulation recognition using support vector machines;

Based on the approximate invariance of Reachability Grammiam on injections of different operating points around normal operating point, the acceleration method combining with the coherency identification method can identify instability modes quickly at the normal operating point by use of the K-medoids algorithm.

They are probabilistic optimal power flow, transfer capability and ancillary services, coherency identification, and large-scale power grid stability analysis.

In order to forecast the unstable modes of an electric power system,a new generator coherency identification approach was proposed by analyzing the state matrix of the system in unstable state.