Open AccessDetect Black Box Signals with Enhanced Spectrum and Support Vector Classifier
Author(s) -
Chenlu Lou,
Xiang Pan
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1438/1/012003
Subject(s) - support vector machine , computer science , narrowband , hyperplane , classifier (uml) , speech recognition , underwater , artificial intelligence , pattern recognition (psychology) , wideband , multipath propagation , acoustics , engineering , channel (broadcasting) , electronic engineering , telecommunications , mathematics , physics , geology , oceanography , geometry
When the plane crash into the sea and the search and rescue team search the location black box, we need to detect a specific acoustic signal in the ocean. The underwater acoustic channel has several distinct features: multipath propagation, high attenuation, and sound velocity varying as a function of depth and temperature, which make the detection of underwater signals very difficult. This paper proposes a new method in the marine signal detection and discrimination based on adaptive line enhancer and support vector machine classifier, which can provide a new idea for marine black box search. The adaptive line enhancer of the least mean square algorithm can detect narrowband signals hidden in wideband noise, while the support vector machine classifier maps signals to high-dimensional space and screens out hyperplanes separating different signals from machine data through previous data. Experimental data show that this method can bring a very high recognition rate.