{"title":"基于似然的宽带DOA子带估计组合","authors":"Petri Helin, B. Dumitrescu, J. Astola, I. Tabus","doi":"10.1109/ISPA.2013.6703760","DOIUrl":null,"url":null,"abstract":"We propose a method for the estimation of direction of arrival (DOA) of wideband sources using a sensor array, especially directed for improving the DOA estimation accuracy in the case of close sources. Starting from the spatial spectrum estimated by MVDR separately in each subband, we select a number of peaks in each subband and build the likelihood of a direction based on the probability distribution of the closest peak. We are resorting to existing bias and variance formulas for MVDR estimates in the narrowband case, accounting for the bias due to multiple sources. The log-likelihood functions are combined additively over the subbands, as opposed to the traditional additive combination of MVDR spatial spectra in each subband. The maxima of the combined log-likelihood function are shown experimentally to be more accurate than the maxima of the combined MVDR spectra.","PeriodicalId":425029,"journal":{"name":"2013 8th International Symposium on Image and Signal Processing and Analysis (ISPA)","volume":"117 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2013-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"2","resultStr":"{\"title\":\"Likelihood based combining of subband estimates for wideband DOA\",\"authors\":\"Petri Helin, B. Dumitrescu, J. Astola, I. Tabus\",\"doi\":\"10.1109/ISPA.2013.6703760\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"We propose a method for the estimation of direction of arrival (DOA) of wideband sources using a sensor array, especially directed for improving the DOA estimation accuracy in the case of close sources. Starting from the spatial spectrum estimated by MVDR separately in each subband, we select a number of peaks in each subband and build the likelihood of a direction based on the probability distribution of the closest peak. We are resorting to existing bias and variance formulas for MVDR estimates in the narrowband case, accounting for the bias due to multiple sources. The log-likelihood functions are combined additively over the subbands, as opposed to the traditional additive combination of MVDR spatial spectra in each subband. The maxima of the combined log-likelihood function are shown experimentally to be more accurate than the maxima of the combined MVDR spectra.\",\"PeriodicalId\":425029,\"journal\":{\"name\":\"2013 8th International Symposium on Image and Signal Processing and Analysis (ISPA)\",\"volume\":\"117 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2013-09-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"2\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2013 8th International Symposium on Image and Signal Processing and Analysis (ISPA)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/ISPA.2013.6703760\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2013 8th International Symposium on Image and Signal Processing and Analysis (ISPA)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ISPA.2013.6703760","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Likelihood based combining of subband estimates for wideband DOA
We propose a method for the estimation of direction of arrival (DOA) of wideband sources using a sensor array, especially directed for improving the DOA estimation accuracy in the case of close sources. Starting from the spatial spectrum estimated by MVDR separately in each subband, we select a number of peaks in each subband and build the likelihood of a direction based on the probability distribution of the closest peak. We are resorting to existing bias and variance formulas for MVDR estimates in the narrowband case, accounting for the bias due to multiple sources. The log-likelihood functions are combined additively over the subbands, as opposed to the traditional additive combination of MVDR spatial spectra in each subband. The maxima of the combined log-likelihood function are shown experimentally to be more accurate than the maxima of the combined MVDR spectra.