摘要
ＡＮｅｕｒａｌＮｅｔｗｏｒｋＭｅｔｈｏｄｆｏｒＭａｔｒｉｘＥｉｇｅｎｓｏｌｖｅｒｉｎＲｅａｌＴｉｍｅＴａｎＹｉｎｇ（谭营）ＨｅＺｈｅｎｙａ（何振亚）（ＤｅｐａｒｔｍｅｎｔｏｆＲａｄｉｏＥｎｇｉｎｅｅｒｉｎｇ，ＳｏｕｔｈｅａｓｔＵｎｉｖｅｒｓｉｔｙ，...
A neural network for eigen decomposition of a positive semi definite matrix is proposed. We show the stability and real time eigen decomposition ability of the proposed neural network. The network can go into its stable state in the magnitude of the circuit time constant. The output voltage of the network under its steady state is just the eigenvector corresponding to the minimum eigenvalue of the NNs connection strength matrix, which is the data covariance matrix without any pre processing. By taking proper processing for R , we can further extract the other eigen parameters of R , sequentially. Both the theoretical analysis and the experimental simulation results show that the proposed network can perform the extraction of the eigenstructure in real time.
