Page 59 - 《应用声学》2019年第6期
P. 59

第 38 卷 第 6 期                                                                       Vol. 38, No. 6
             2019 年 11 月                         Journal of Applied Acoustics                 November, 2019

             ⋄ 研究报告 ⋄



                  流形学习在运动声源声特征提取方面的研究                                                                ∗





                           宿元亮      1,2  刘志红      1,2  王万凯      1,2  赵玉贵      1,2  仪垂杰     1,2†



                                          (1  青岛理工大学机械与汽车工程学院         青岛   266000)
                                       (2  工业流体节能与污染控制教育部重点实验室           青岛   266000)

                摘要    运动声源因声信号时变性、叠加性和空时耦合性强，声数据呈现高维、非线性等特点，使得关键声特征
                提取困难，声特征提取方法复杂度高、数值计算量大、有效性差。因此，如何有效提取声特征并降低提取方法复
                杂度成为目前多源声场声源精准识别需迫切解决的关键科学问题。由此，该文提出短时傅里叶变换 (STFT)
                和局部线性嵌入算法 (LLE) 联合的 STFT-LLE 流形学习声特征提取方法，并将此方法应用于运动声特征提
                取，且通过仿真实验测试对其进行了验证。该方法为运动声目标的分类识别提供了技术支撑。
                关键词     运动声源，特征提取，流形学习，短时傅里叶变换，局部线性嵌入
                中图法分类号: X827           文献标识码: A          文章编号: 1000-310X(2019)06-0961-08
                DOI: 10.11684/j.issn.1000-310X.2019.06.008




                   Research on manifold learning in acoustic feature extraction of moving
                                                     sound source




                   SU Yuanliang  1,2  LIU Zhihong  1,2  WANG Wankai  1,2  ZHAO Yugui 1,2  YI Chuijie  1,2

                  (1 College of Mechanical and Automotive Engineering, Qingdao University of Technology, Qingdao 266000, China)
                    (2 Key Laboratory of Energy Conservation and Pollution Control of Industrial Fluids, Ministry of Education,
                                                    Qingdao 266000, China)

                 Abstract  Moving noise is characterized by time-varying, superimposing and space-time coupling of sound
                 signals, and the sound data are characterized by high dimensionality and nonlinearity, which makes it difficult
                 to extract key acoustic features. The method of sound feature extraction has high complexity, large numerical
                 calculation and poor validity. Therefore, how to effectively extract acoustic features and reduce the complexity
                 of the extraction method has become an important scientific problem for the accurate identification of multi-
                 source acoustic sources. In this paper, the STFT-LLE manifold learning method is proposed. It combined with
                 short-time Fourier transform (STFT) and local linear embedding algorithm (LLE). This method is applied to
                 the feature extraction of motion acoustic field. It is validated by simulation experiments. This method provides
                 technical support for the classification and recognition of moving sound source.
                 Key words Moving sound source, Feature extraction, Manifold learning, Short time Fourier transform,
                 Locally linear embedding


             2019-02-16 收稿; 2019-09-11 定稿
             国家自然科学基金项目 (61671262, 61871447 )
             ∗
             作者简介: 宿元亮 (1993- ), 男, 山东泰安人, 硕士研究生, 研究方向: 噪声与振动控制和声信号处理。
             † 通讯作者 E-mail: chuijieyi@vip.163.com
   54   55   56   57   58   59   60   61   62   63   64