Search

In order to achieve anti-jamming, radar needs to identify the jamming signals to distinguish between the target echo signals and the jamming signals. Interrupted sampling repeater jamming (ISRJ) is a widely used jamming type in radar and electronic countermeasures. In this paper, we propose an efficient channel attention (ECA)-based one-dimensional residual neural network method for the ISRJ signal recognition. The joint features based on axially integral bispectrum and square integral bispectrum are used to extract the target echo signal and ISRJ signal. The ECA-based one-dimensional residual neural network is applied to the ISRJ recognition. The experiment results show that the proposed ECA-based one-dimensional residual neural network has better correct recognition rate under low signal-to-noise ratio and can offer 30.17% reduction in timeliness than the improved convolutional neural network-residual network (CNN-ResNet) method. [ABSTRACT FROM AUTHOR]

Software radio channelizer is employed in the multi-standard wireless communication for extracting individual channels from the wideband input signal. Frequency response masking (FRM) finite impulse response (FIR) filters are used in software radio channelizer for reducing the computational complexity of the maximally decimated filter bank (MDFB) with narrow transition band (NTB). In this paper, we present an approach to design two well-known classes of efficient digital filters, namely the interpolated and full FRM FIR filters. In the proposed MDFB, the FRM is used to obtain flexible transition band and low complexity. The proposed improved FRM-based MDFB with NTB can be directly applied to high sampling rate software radio channelizer. We show the FRM filter examples for single carrier orthogonal frequency division multiplexing (SC-OFDM) modulator and wideband digital receiver (WDR). The simulation results indicate that the improved FRM-based MDFB with NTB is correct. Design examples with complexity comparison are presented to demonstrate the complexity improvement of the proposed FRM-based MDFB with NTB. [ABSTRACT FROM AUTHOR]