碩士 / 中原大學 / 資訊工程學系 / 87 / In the process of spreading various underwater acoustic signals in the ocean, because of being affected by the ocean environment and all kinds of oceanic noises permeate through, it is necessary to have a appropriate signal procedure after receiving to identify the signals that energy is already decreased by long distance transmission and environment interruption. To study real-time underwater acoustic signal identification, it is divided into two important parts: the first research subject is wavelet analysis using in signal feature selection; the second is the identify application of fuzzy logic. Combine two parts and establish a reality, low cost, real-time identify system.
When selecting feature parameter analysis, study signal character analysis and how to get the feature parameter individually. Being proved, in terms of the character of wavelet analysis multi-resolution and study different feature parameter selective method can get the feature parameter that realistic react different ships' character. Besides, use self-organization neural network training stage to clustering data, and find the character of data gathering center can get representative pattern feature parameter of each sample classification individually.
Use fuzzy logic theory to establish identification system and make fuzzy logic rules by various classified pattern features. Get the identifiable result after using fuzzy logic recognition system.
To the purpose of real-time identification, use multi-task and dual- buffer mode can let the signal collection and identification concurrent run which can improve the performance efficiency of the system, and establish a realistic real-time acoustic signal identify system. In addition, structure real-time client-server identify simulation system on the network can let the development of identify system more elastic.
| Identifer | oai:union.ndltd.org:TW/087CYCU0392023 |
| Date | January 1999 |
| Creators | Liang-Ching Lin, 林良清 |
| Contributors | Cku-Kuei Tu, 杜筑奎 |
| Source Sets | National Digital Library of Theses and Dissertations in Taiwan |
| Language | zh-TW |
| Detected Language | English |
| Type | 學位論文 ; thesis |
| Format | 70 |
Page generated in 0.0061 seconds