基于迭代中心差分卡尔曼滤波的说话人跟踪方法

侯代文; 殷福亮

doi:10.3724/SP.J.1146.2006.01897

基于迭代中心差分卡尔曼滤波的说话人跟踪方法

doi: 10.3724/SP.J.1146.2006.01897

侯代文^{①②;殷福亮},
殷福亮

基金项目:

国家自然科学基金(60372082)和教育部跨世纪优秀人才基金资助课题

计量
- 文章访问数: 3334
- HTML全文浏览量: 94
- PDF下载量: 822
- 被引次数: 0
出版历程
- 收稿日期: 2006-11-30
- 修回日期: 2007-06-08
- 刊出日期: 2008-07-19

Iterated Central Difference Kalman Filter Based Speaker Tracking

Hou Dai-wen^{①②;殷福亮},
Yin Fu-liang

摘要

摘要: 利用状态空间方法对说话人进行语音跟踪时,观测方程的非线性会影响说话人位置的估计精度。该文将迭代滤波理论与中心差分卡尔曼滤波技术相结合,提出迭代的中心差分卡尔曼滤波方法,并应用于说话人跟踪系统。仿真实验结果表明,该文所提出的方法减少了系统线性化误差,增强了滤波算法的鲁棒性,提高了说话人跟踪精度。
- 说话人跟踪; 中心差分; 卡尔曼滤波; 麦克风阵列; Levenberg-Marquardt方法
Abstract: In the state space method based speaker tracking system, the nonlinearity of the measurement function degrades the localizing accuracy of the speaker tracking method severely. The iterated central difference Kalman filter algorithm, which incorporates the iterated filtering theory and the Central Difference Kalman Filter (CDKF) method, is proposed to reduce linearization error. In comparison with traditional CDKF method, the proposed method has higher tracking accuracy, faster convergence speed and more robust stability. Simulation results demonstrate the effectiveness of the proposed method.

HTML全文

参考文献(1)

Potamitis I, Chen H, and Tremoulis G. Tracking of multiplemoving speakers with multiple microphone arrays[J].IEEETrans. on Speech and Audio Processing.2004, 12(5):520-529[2]Bangs W and Schultheis P. Space-time processing for optimalparameter estimation. in Signal Processing, Grriffiths J,Stocklin P and Schooneveld C V Eds., New York: AcademicPress , 1973: 577-590.[3]Brandstein M A. Framework for speech source localizationusing sensor arrays. [Ph.D. Thesis], Brown University,Providence, RI, U.S.A, 1995.[4]Sturim D, Brandstein M, and Silverman H. Trackingmultiple talkers using microphone array measurements.Proceedings of IEEE International Conference on Acoustics,Speech and Signal Processing, Munich, Germany, 1997:371-374.[5]Bizwas K K and Mahalanabis A K. Suboptimal algorithm fornonlinear smoothing[J].IEEE Trans. on Aerospace andElectronic Systems.1973, 9(4):529-534[6]Julier S and Uhlmann J K. Unscented filtering and nonlinearestimation[J].Procee. IEEE.2004, 92(3):401-422[7]Ngarrd M, Poulsen N, and Ravn O. New developments instate estimation for nonlinear systems[J].Automatica.2000,36(11):1627-1638[8]Ito K and Xiong K. Gaussian filters for nonlinear filteringproblems[J].IEEE Trans. on Automatic Control.2000, 45(5):910-927[9]Dennis J and Schnabel R. Numerical Methods forUnconstrained Optimization and Nonlinear Equations.Philadelphia: SIAM, 1996: 218-236.[10]Bar-Shalom Y and Fortmann T. Tracking and DataAssociation. London: Academic Press Inc, 1988: 119-122.[11]Bellaire R L, Kamen E W, and Zabin S M. A new nonlineariterated filter with applications to target tracking. inProceedings of 1995 SPIE Conference on Signal and DataProcessing of Small Targets, Orlando, FL, USA, April 1995,2561: 240-251.[12]More J. The Levenberg-Marquardt algorithm: implementationand theory. in Numerical Analysis, Watson G A Ed., LectureNotes in Math, 630, Springer Verlag, Berlin: 1978: 105-116.[13]Vermaak J and Blake A. Nonlinear filtering for speakertracking in noisy and reverberant environments. IEEEInternational Conference on Acoustics, Speech, and SignalProcessing, Salt Lake City, USA, 2001: 3021-3024.[14]Knapp C H and Carter G C. The generalized correlationmethod for estimation of time delay[J].IEEE Trans. onAcoustics, Speech and Signal Processing.1976, 24(4):320-327[15]Allen J B and Berkley D A. Image method for efficientlysimulating small-room acoustics[J].Journal of the AcousticalSociety of America.1979, 65(4):943-950

施引文献

资源附件(0)

访问统计