Broadband Direction of Arrival Estimation Based on Sparse Least Squares Support Vector Machines
Zhong Zi-fa · Journal of Chinese Computer Systems · 2012
Broadband direction finding characteristic of antenna is very complex,so the efficient learning algorithm and the optimal training sets are very important when using smart learning method to estimate the direction of arrival.The paper use LS-SVM algorithm to build DOA estimation model which can solve the problem faster,prune the non support vectors of the LS-SVM,then construct the DOA estimation model use high support vectors as training sets for second learning.Experimental results show that the second learning with the sparse vectors can prove the DOA precision,and has a broad application value in broadband DOA estimation.