Bayesian beamforming based on probabilistic signal model in uncertain oceanic environments
Even a determined signal is transmitted, the received signal propagating through an uncertain oceanic environment can’t be predicted exactly. Instead, it should be regard as a stochastic acoustic field. The probabilistic signal model, is a robust received signal model, which describes the received signal fidelity and completeness. The model predicts the statistical characteristics of the received signal by embedding the oceanic environmental uncertainty into the propagating model, includes the dominant signature represented by the mean field and the uncertainty by the variability. One method that incorporates the environmental variability into acoustic field is polynomial chaos expansion method. The statistical characteristics of the received signal are then used to detect and localize targets. Sonar signal processing can be generalized as beamforming which is a quadratic operation of propagating signal replica vector with covariance matrix of array data. The replica vector is obtained by the propagation model, while the covariance matrix is estimated from data. Based on the probabilistic signal model, a Bayesian beamforming is derived to against the mismatch of signal replica vector. The Bayesian beamforming turns out an estimator-correlator structure , which combines a prior of model with data knowledge effectively, and hence realizes robust detecting and estimating of a target.
ZHAO HangFang GONG XianYi YU ZiBin
Hangzhou Applied Acoustics Research Institute,Hangzhou City,Zhejiang Province,China Department of In Hangzhou Applied Acoustics Research Institute,Hangzhou City,Zhejiang Province,China
国际会议
The 10th Western Pacific Acoustics Conference(第十届西太平洋声学会议WESPAC X)
北京
英文
1-10
2009-09-21（万方平台首次上网日期，不代表论文的发表时间）