Dissertation > Excellent graduate degree dissertation topics show
Extraction and classification recognition of ship radiated noise characteristics based on higher-order statistics
Author: ZhangYiJun
Tutor: WangHaiYan
School: Northwestern Polytechnical University
Course: Mechanical and Electronic Engineering
Keywords: Higher order cumulants Order cyclic cumulants 1 ( 1 / 2 ) -dimensional spectrum 2 ( 1 / 2 ) -dimensional spectral phase coupling B-P neural networks Target Identification and Classification
CLC: TP14
Type: Master's thesis
Year: 2001
Downloads: 401
Quote: 7
Read: Download Dissertation
Abstract
|
Higher-order statistics is an effective tool to study the nonlinear and non - Gaussian signal , it has unique advantages in signal detection , feature extraction and harmonic retrieval . This paper application of higher-order statistics feature extraction and classification and identification of underwater targets . The paper first analyzes the physical properties of the ship radiated noise , made ??a variety of extraction methods based on the the ship noise characteristic parameters of the higher-order statistics : the bispectrum and trispectrum analysis , 1 ( 1 / 2 ) -dimensional spectral spectrum analysis , 2 ( 1 / 2) -dimensional spectral phase coupling characteristics analysis , and the first order cyclic cumulant method used in ship radiated noise modulation signal feature extraction: the third-order cycle cumulant diagonal slice spectrum extracted noise periodic component . Line continuous spectral characteristics extracted by the above method as characteristic parameters , using BP neural network to identify the classification of the three types of ship noise sample , the average correct recognition rate reached 91.5% . And verify the correctness and validity of the feature extraction using higher-order statistics . This paper studies suggest that the ship noise modulation information contains a large number of ships, the characteristic parameters ( speed, host type , operating conditions , etc.) , the use of high-end cycle cumulative amount extracted modulation information with other methods ( power cepstrum ) compared has many advantages , such as the inhibition of any smooth ( non-stationary ) Gaussian noise, a separable smooth and non-stationary signals etc. , and to get a better identification of the classification results . As the papers workload , this identification classification , denoising and feature parameters selected research needs to be further explored.
|
Related Dissertations
- Research on Anti-periodic Solutions of Delayed Cellular Neural Networks without Assuming Global Lipschitz Conditions,TP183
- Designs and Applications of Fuzzy Synthetic Evaluation Models Based on Parallel Algorithms,TP18
- Research on the Models of GPS Height Fitting Based on BP Neural Network and Their Applications,P228.4
- Based on Artificial Neural Network Classification of Remote Sensing Images,P237
- Researches on the Theories and Algorithms of Earth Orientation Paramters Prediction,P22
- Comprehensive Research on Intellective Deformation Analysis Method,P209
- Theoretical and Experimental Studies on the Measurement of COD in Water Using Ultraviolet Spectrum Method,X832
- RTK-GPS Strategy Based on Genetic Neural Network Digital for Forest Mapping and Positioning of Samples and Its Application,P228.4
- Research on Digital Proofing Examining of Press-oriented Multi-source Map Data,P288
- High Speed Frequency Measurment and Non-Linearity Correction of Frequency Modulated Capacitive Displacement Sensor,TH822
- Use Machine-made Sand to Produce High Performance Concrete in Low Ratio of Water-binder and the Performance’s Nonlinear Prediction,TU528
- Research on Mine Gas Geology in Magmatic Incursion Area,TD712
- Signal Recognition and Parameter Estimation for Multi-carrier Systems in Shortwave Communications,TN919.3
- Research on Interference Type Recognition in DSSS Communication Systems,TN911.7
- The application of data mining technology in the analysis of the measured data,TP311.13
- Research of Unified Power Quality Conditioner Based on Neural Network,TM933.4
- Based on CT images of peripheral non -small cell lung cancer diagnosis of type Model,TP391.41
- The Image Fusion Technology Based on Multiscale Geometric Analysis,TP391.41
- Cellular Neural Networks Based on Variation of Partial Differential Equations Committed to Underwater Image Noise Reduction,TP391.41
- The Research of the Stability of the Neural Network Based on the Nonsmooth Anaslysis,TP183
- The Study of Navigation System of AGV,TP242
CLC: > Industrial Technology > Automation technology,computer technology > Automated basic theory > Automatic Information Theory
© 2012 www.DissertationTopic.Net Mobile
|