PROJECT TITLE :

Integrated Optimization of Long-Range Underwater Signal Detection, Feature Extraction, and Classification for Nuclear Treaty Monitoring

ABSTRACT:

We tend to designed and jointly optimized an integrated signal processing chain for detection and classification of long-range passive-acoustic underwater signals recorded by the global geophysical monitoring network of the great Nuclear-Test-Ban Treaty Organization. Beginning at the amount of raw waveform data, a processing chain of signal detection, feature extraction, and signal classification was designed and jointly optimized to the task. Relevant waveform segments were during a initial step identified by a generic, flexibly parameterized detection algorithm on an extended- to short-term averages' ratio of the spectral energy. For representation, general-purpose sound processing options, with an extra focus on spectral and cepstral features, were extracted from the detected segments. As classifiers, support vector machines with completely different kernel functions were employed alongside other baseline learning algorithms. The free parameters of the general toolchain (i.e., trigger algorithm parameters and classifier hyperparameters) were jointly optimized in an exceedingly cross-validation setting, either in step with the cross-validation classification error or the cross-validation space underneath the receiver operating characteristic curve. Experiments demonstrate that our technique outperforms machine learning algorithms task-tailored to a previous, human-expert-designed preprocessing chain. The presented approach can be tailored to a wide selection of problems that may benefit from jointly optimizing parameters of preprocessing and classification algorithm.