摘要
Aiming at source number determination and direction of arrival(DOA) estimation under the case of time-varying source number,a method of DOA estimation with an unknown number of sources was proposed.Firstly,an algorithm based on crossvalidation technique was introduced to determine the number of sources.Then dynamic DOAs of source were estimated using an algorithm based on blind source separation(BSS) under the case that number of sources were unknown in advance and it was timevarying.The effectiveness of the proposed method was validated by simulation of time-invariant and time-varying numbers of source.Compared with other conventional methods,the proposed method has superior evaluation performances The proposed method can estimate m(the numbers of sensor) DOAs while other conventional methods estimate less than m DOAs.The R_(mse) of the proposed method in the case of low signal-to-noise ratio(SNR)(equal or lower than 30 dB) is smaller than 0.2 while R_(mse) of other conventional methods are greater than 0.8.
Aiming at source number determination and direction of arrival（DOA） estimation under the case of time-varying source number,a method of DOA estimation with an unknown number of sources was proposed.Firstly,an algorithm based on crossvalidation technique was introduced to determine the number of sources.Then dynamic DOAs of source were estimated using an algorithm based on blind source separation（BSS） under the case that number of sources were unknown in advance and it was timevarying.The effectiveness of the proposed method was validated by simulation of time-invariant and time-varying numbers of source.Compared with other conventional methods,the proposed method has superior evaluation performances The proposed method can estimate m（the numbers of sensor） DOAs while other conventional methods estimate less than m DOAs.The R_（mse） of the proposed method in the case of low signal-to-noise ratio（SNR）（equal or lower than 30 dB） is smaller than 0.2 while R_（mse） of other conventional methods are greater than 0.8.
基金
National Natural Science Foundation of China(No.51309116)
the Foundation of Fujian Education Committee for Distinguished Young Scholars,China(No.JA14169)
the Scientific Research Foundations of Jimei University,China(Nos.ZQ2013001,ZC2013012)
Open Project of Artificial Intelligence Key Laboratory of Sichuan Province,China(No.2014RYJ03)
Natural Science Foundation of Fujian Province,China(No.2016J01736)
