Tracking Direction Selection Method Using Automatic Associative Memory Neural Network

摘要

The present invention relates to a tracking direction selection method using an ASM neural network that can reduce the tracking error caused by instantaneous changes due to detection threshold (DT) setting and background noise using an automatic associative memory (ASM) neural network. N-channel underwater signal is collected by using the linear array sensor. A first step of performing an FFT on the second switch, a second step of forming a cardioid beam in order to solve the symmetry of the target appearance orientation of the linear array sensor of the first step, and a cardioid beam formed in the second step. A third step of extracting an energy sum and a tonal sum using a frequency characteristic for each azimuth, a fourth step of removing background noises of the energy sum and a tonal sum extracted in the third step using a two pass min method; After performing the fourth step, the DT is selected by using the automatic associative memory device, and the fifth step of selecting the tracking direction by allocating all the above-mentioned defense bins to 1 and assigning the bearing bins below the DT to 0 respectively. By using the tracking algorithm using ASM neural network instead of the α-β tracking algorithm, it is possible to reduce the DT selection error and tracking direction selection error due to background noise when selecting the tracking direction in the passive sonar It is effective to be.