2017 Fiscal Year Final Research Report
Optimal variable step-size LMS algorithm and application to vibration analysis
Project/Area Number |
15K06117
|
Research Category |
Grant-in-Aid for Scientific Research (C)
|
Allocation Type | Multi-year Fund |
Section | 一般 |
Research Field |
Measurement engineering
|
Research Institution | Prefectural University of Hiroshima |
Principal Investigator |
Xiao Yegui 県立広島大学, 経営情報学部, 教授 (50252325)
|
Project Period (FY) |
2015-04-01 – 2018-03-31
|
Keywords | 適応フーリエ解析 / LMSアルゴリズム / 可変ステップサイズLMS / 変分法 / 適応IIRノッチフィルタ / 振動解析 / 狭帯域能動騒音制御 / 統計解析 |
Outline of Final Research Achievements |
Adaptive Fourier analyzer has been considered to be a useful and practical tool which can overcome the two well-known major difficulties (frequency resolution and tracking capability) with the traditional FFT. The LMS algorithm is much preferred in real-life applications, but the tradeoff between its convergence and steady-state performance has been a big issue. In recent years, the variable step-size LMS (VSS-LMS) algorithm has attracted a lot of attention. In this project, after carefully investigating the existing VSS-LMS algorithms, we introduced a general VSS update equation and used calculus of variations to optimize it. As a result, several VSS-LMS algorithms have been proposed and confirmed to present promising performance and might be applied to real vibration analysis, active noise control, etc. with a very moderate increase in computational cost as compared to the conventional VSS-LMS algorithm.
|
Free Research Field |
知能信号処理
|